2012 Book TheFutureInternet

Lecture Notes in Computer Science 7281
Commenced Publication in 1973

Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David Hutchison
Lancaster University, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Alfred Kobsa
University of California, Irvine, CA, USA
Friedemann Mattern
ETH Zurich, Switzerland
John C. Mitchell
Stanford University, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
Oscar Nierstrasz
University of Bern, Switzerland
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, Germany
Madhu Sudan
Microsoft Research, Cambridge, MA, USA
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbruecken, Germany
Federico Álvarez Frances Cleary
Petros Daras John Domingue
Alex Galis Ana Garcia
Anastasius Gavras Stamatis Karnourskos
Srdjan Krco Man-Sze Li
Volkmar Lotz Henning Müller
Elio Salvadori Anne-Marie Sassen
Hans Schaffers Burkhard Stiller
Georgios Tselentis Petra Turkama
Theodore Zahariadis (Eds.)
The Future Internet
Future Internet Assembly 2012:

From Promises to Reality
13
Volume Editors
Federico Álvarez Stamatis Karnourskos Hans Schaffers

Frances Cleary Srdjan Krco Burkhard Stiller
Petros Daras Man-Sze Li Georgios Tselentis
John Domingue Volkmar Lotz Petra Turkama
Alex Galis Henning Müller Theodore Zahariadis
Ana Garcia Elio Salvadori
Anastasius Gavras Anne-Marie Sassen
Acknowledgement and Disclaimer

The work published in this book is partly funded by the European Union under the
Seventh Framework Programme. The book reflects only the authors’views. The Union
is not liable for any use that may be made of the information contained therein.
ISSN 0302-9743 e-ISSN 1611-3349

ISBN 978-3-642-30240-4 e-ISBN 978-3-642-30241-1
DOI 10.1007/978-3-642-30241-1
Springer Heidelberg Dordrecht London New York
Library of Congress Control Number: 2012937591
CR Subject Classification (1998): C.2, D.4.4, D.2, H.3.5, H.4, K.6.5
LNCS Sublibrary: SL 3 – Information Systems and Applications, incl. Internet/Web

and HCI
© The Editor(s) (if applicable) and the Author(s) 2011. The book is published with open access at Springer-
Link.com
Open Access. This book is distributed under the terms of the Creative Commons Attribution Noncommercial
License which permits any noncommercial use, distribution, and reproduction in any medium, provided the
original author(s) and source are credited.
This work is subject to copyright for commercial use. All rights are reserved, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other way, and storage in data banks. Duplication of
this publication or parts thereof is permitted only under the provisions of the German Copyright Law of
September 9, 1965, in its current version, and permission for use must always be obtained from Springer.
Violations are liable to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, 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.
Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
List of Editors
Federico Álvarez
Universidad Politécnica de Madrid, Spain
[email protected]
Frances Cleary
Waterford Institute of Technology - TSSG, Waterford, Ireland
[email protected]
Petros Daras
CERTH-ITI, Thessaloniki, Greece
[email protected]
John Domingue
Knowledge Media Institute, The Open University,
Milton Keynes, UK
[email protected]
Alex Galis
Department of Electronic and Electrical Engineering,
University College London, UK
[email protected]
Ana Garcia
ENoLL, Brussels, Belgium
[email protected]
Anastasius Gavras
Eurescom GmbH, Heidelberg, Germany
[email protected]
Stamatis Karnouskos
SAP Research, Karlsruhe, Germany
[email protected]
Srdjan Krco
Ericsson Serbia, Belgrade, Serbia
[email protected]
VI List of Editors
Man-Sze Li
IC Focus, London, UK
[email protected]
Volkmar Lotz
SAP Research, Sophia Antipolis, France
[email protected]
Henning Müller
Business Information Systems, University of Applied Sciences Western
Switzerland, Sierre, Switzerland
[email protected]
Elio Salvadori
CREATE-NET, Trento, Italy
[email protected]
Anne-Marie Sassen
European Commission, Brussels, Belgium
[email protected]
Hans Schaffers
ESoCE Net, Dialogic, Aalto University, Aalto, Finland
[email protected]
Burkhard Stiller
University of Zürich, Switzerland
[email protected]
Georgios Tselentis
European Commission, Brussels, Belgium
[email protected]
Petra Turkama
CKIR, Aalto University, Aalto, Finland
[email protected]
Theodore Zahariadis
Synelixis/TEI of Chalkida, Greece
[email protected]
Preface
This publication constitutes the 2012 edition of the yearly Future Internet
Assembly book, which has been published since 2009.
The Future Internet Assembly (FIA) is a successful, unique, and bi-annual
conference that brings together participants of over 150 projects from several
distinct but interrelated areas in the EU Framework Programme 7.
They share scientific and technical results and discuss cross-domain research
topics around the notion of creating new Future Internet technologies, applica-
tions, and services with a global view.
FIAs history started in spring 2008 in Bled, Slovenia, and the spring of 2012
saw the 9th FIA conference in Aalborg, Denmark. As with prior spring FIAs,
the community has put together a book, which aggregates both representative
results achieved in the Future Internet domain and the possibilities of what can
be expected in the medium or short term.
In the FIA time line several key elements were required to ensure success.
These are:
- Cross-domain considerations: on both core technical issues, such as FI ar-
chitectures, FI services, FI experimentation, mobile FI, or Internet of Things,
and on horizontal issues, such as socio-economics, privacy, trust, and identity.
- Engagement with application areas of the Future Internet and users: to move
from FI technologies to sectors where innovation can be improved by Future
Internet technologies.
- Provision of results that are applicable in day-to-day life.
Within the structure of the book, different topics are covered in a balanced
and coherent manner.
The topics of the book have been organized into four chapters:
- Future Internet foundations cover core cross-domain technical and horizon-
tal topics. Chapters within this section include architectural questions; mobile
Internet, cloud computing, socio-economic questions; trust and identity; search
and discovery; and experiments and experimental design.
- Future Internet technical areas are those technical domains that are as-
sociated to the Future Internet, mainly but not limited to networks, services,
Internet of Things, content, and cross-area questions.
- Future Internet application areas consist of user areas and communities
where the Future Internet can boost innovation. The chapters within this sec-
tion cover smart cities, smart energy, smart health, smart enterprises, smart
environment, smart transportation, logistics and mobility, smart manufacturing,
smart agriculture, and tourism.
- Future Internet infrastructures cover experimentation and results in real
infrastructures within the FI domain.
VIII Preface
There were 40 submissions. Each submission was peer-reviewed by experts

in the field and editors of the book. The committee decided to accept 20 pa-
pers. Introductions to the four chapters of the book and an invited introduction
describing the FIA Roadmap are also provided.
We would like to acknowledge the hard work of the reviewers of the book, and
the support provided by Easychair, which was used for the electronic submission
and paper review.
Last but not least we would like to mention the European FP7 projects
that financially supported the book publication: FIRESTATION, EFFECTS+,
SESERV, and CONCORD (on behalf of the FI-PPP).
March 2012 Federico Álvarez

Organization
Programme Committee
Federico Álvarez Universidad Politécnica de Madrid
Frances Cleary Waterford Institute of Technology - TSSG
Petros Daras Informatics and Telematics Institute, CERTH
John Domingue Knowledge Media Institute, The Open
University
Alex Galis University College London
Ana Garcia ENoLL
Anastasius Gavras Eurescom GmbH
Stamatis Karnourskos SAP Research
Srdjan Krco Ericsson
Man-Sze Li IC Focus
Volkmar Lotz SAP AG
Henning Müller University of Applied Sciences Western
Switzerland, Sierre (HES-SO)
Elio Salvadori CREATE-NET
Anne-Marie Sassen European Commission
Hans Schaffers Aalto University
Burkhard Stiller University of Zurich
Georgios Tselentis European Commission
Petra Turkama Aalto University
Theodore Zahariadis Synelixis Ltd.
Additional Reviewers
Apostolos Axenopoulos Jacek Kopecky
Iacopo Carreras Ning Li
Smitashree Choudhury Daniele Miorandi
Dejan Drajic Nicu Sebe
Antonio Foncubierta Domenico Siracusa
Nenad Gligoric Dimitris Zarpalas
Vlastuin Group
Future Internet Foundations
Introduction
Burkhard Stiller
University of Zürich, CSG@IFI, Zürich, Switzerland
Alex Galis
University College London, U.K.
Theodore Zahariadis
Synelixis Ltd.
Since the Internet has reached a major standing in terms of being the medium
and the infrastructure for information exchange, management and computation
in the beginning of the 21st century, its usage is characterized by billions of users
as well as by hundreds, even thousands of different applications and services.
The diversity of these applications and the use of today’s Internet outline a very
successful approach and undisputedly determine the core pillar of the Informa-
tion and Communication Technology (ICT) landscape and its societal challenges
and opportunities. Such a success is deemed naturally positive, however, the
scale of a distributed system typically determines as well the overall achievable
performance, a degree of user satisfaction, and operational perspectives. There-
fore, the operational and commercial dimensions of Internet communications and
computation have turned into areas, which go well beyond the initial Internet’s
technology and its basics, including areas such as high reliability, full-fledged
security, mobility support, or delay tolerance. While these technology-driven di-
mensions are being enriched by application-specific, provider-critical, and user-
driven facets, the set of economic and societal factors of major importance are
being addressed in work in the context of Future Internet, too. Thus, the need to
re-think at least partially the Future Internet foundations is essential, especially
to enable future networks and services as well as novel technology to be able to
cope with those new demands.
In consequence, the addressing of relevant, important, and arising foundations
of a Future Internet are crucial for a success of new infrastructures to come. As
such the pure delivery of packets - one of the key design principles for a robust
Internet – has to be extended with those principles, which have to guide future
developments. In addition, the analysis of technology-to-economic relations in
terms of inter-stakeholder operations is essential for a modern Future Internet
Foundation, as the economic dimension of the information exchange has reached
it technical limitations of today’s Internet.
A particular key aspect is the study of system limits defining the constraints
and freedoms in controlling the Future Internet. Limits can be determined by
analyzing how the behaviour of the system depends on the parameters that drive
the system. Some limits would lead to unexpected and significant behaviour
changes of the system, for example the unpredictable boundaries or changes
XII Future Internet Foundations
in the scale of magnitude. Some other limits are determined by non-common

behaviour interactions between the components of a system.
And as highly distributed and decentralized systems can interwork only, if
the right standard is in place, the important paths to a systematic research to
standardization approach are required. To address a third domain of relevance
for the Future Internet (besides the technology itself and its economics) the so-
cietal aspect determines a highly relevant cross-disciplinary facet. In that sense
the relations between technology, pervasiveness, and societal foundations require
a carefully study and guidelines’ work. Last but not least a number of detailed
technical advances play an important role for Future Internet generations, such
as combined development and run-time environments or interactive and the an-
alytics of interaction and security solutions in an Internet.
Therefore, the content of the book section on “Future Internet Foundations”
addresses selected topics in this field, and it could be classified as follows:
• Future Internet Basics: Design principles and tussle analysis
• Future Internet Standardization: Systematic approach and development path
• Future Internet Cross-disciplinary Facets: Technology, pervasiveness, and
society
• Future Internet Technology Advances: Development and run-time environ-
ments as well as interaction and security solutions.
Following the review and selection process run for this FIA book, eight chap-
ters were chosen covering some of the above research in Future Internet with
two papers in each category. The following is a summary of main results of the
“Future Internet Foundations” section of this FIA book.
The paper on “Design Principles for the Future Internet Architecture” by
Dimitri Papadimitriou, Theodore Zahariadis, Pedro Martinez-Julia, Ioanna
Papafili, Vito Morreale, Francesco Torelli, Bernard Sales, and Piet Demeester
addresses very basic foundations of systems: design principles. Here, the authors
consider design principles being a central role in the architecture of the Internet,
which have driven and will drive most engineering decisions at a conception level
and operational level. While the paper’s content is based on the EC Future In-
ternet Architecture (FIArch) Group results, it identifies those design principles,
which are expected to govern the future architecture of the Internet.
The work by Alexandros Kostopoulos, Ioanna Papafili, Costas Kalogiros,
Tapio Levä, Nan Zhang, and Dirk Trossen on“A Tussle Analysis for Information-
centric Networking Architectures” highlights in an examination, based on the
tussle analysis method, key interests of various stake-holders, which shall to
be taken into account by future designers when deploying new content deliv-
ery schemes under the Information Centric-Network (ICN) paradigm. This is
considered highly relevant, since key concepts of ICNs are expected to have sig-
nificant impact on the Future Internet, especially by creating new challenges for
all associated stakeholders.
The approach by Bernard Sales, Emmanuel Darmois, Dimitri Papadimitriou,
and Didier Bourse on “A Systematic Approach for Closing the Research to Stan-
dardization gap” argues that standardization activities are recognized as one of
Future Internet Foundations XIII
the tools to incubate research results and accelerate their transfer to innovative
marketable products and services. But since a lack of research transfer via the
standardization channel is visible in EU research, generally referred to as the
research-to-standardization gap, this paper analyzes the root causes for this sit-
uation and proposes a research-focused pre-standardization as a supplemented
methodology and its associated processes to aim at a systematic analysis of
standardization aspects of research projects.
The paper on “From Internet Architecture Research to Standards” by Dimitri
Papadimitriou, Bernard Sales, Piet Demeester, and Theodore Zahariadis argues
that that the debate between architectural research driven by the application
of the theory of utility and the theory of change is over. It highlights a “third
path” which is based on identifying the actual foundational design principles of
the Internet such as the modularization principle and by acknowledging the need
for a all-inclusive architecture instead of (re-) designing protocols independently
and expecting that their combination would lead to a consistent architecture
at running time. The proposed path will in turn also partially impact how the
necessary standardization work is to be organized and conducted, including both
“problem-driven” and “architecture driven” work.
The work on “SOCIETIES: Where Pervasive Meets Social ” by Kevon Doolin,
Ioanna Roussaki, Mark Roddy, Nikos Kalatzis, Elizabeth Papadopoulou, Nick
Taylor, Nicolas Liampotis, David McKitterick, Edel Jennings, and Pavlos Kos-
mides provides an overview of the vision, concepts, methodology, architecture,
and initial evaluation of results toward the accomplishment of the goal to im-
prove the utility of Future Internet services by combining benefits of pervasive
systems with those of social computing. As such, the work in the SOCITIES In-
tegrated project attempts to bridge different technologies in a unified platform,
especially by allowing individuals to utilize pervasive services in a community
sphere.
The lessons learned on “Cross-Disciplinary Lessons for the Future Internet”
by Anne-Marie Oostveen, Isis Hjorth, Brian Pickering, Michael Boniface, Eric
T. Meyer, and Cristobal Cobo are described in terms of socio-economic barriers
related to the Future Internet. As the authors outline, these observations are
derived from an on-line survey and a workshop organized by the Coordination
and Support Action SESERV, which identified six key social and economic issues
to be deemed most relevant by 98 representatives from FP7 Challenge 1 projects.
Thus, the cross-disciplinary views (including social scientists, economists, policy
experts, and other stakeholders) are expressed and seen by the Future Internet
community itself. In turn, the paper presents strategies for some solutions to
these challenges, which is complemented by an investigation on how relevant the
European Digital Agenda is to Future Internet technologists.
The view on “An Integrated Development and Runtime Environment for the
Future Internet ”, expressed by Amira Ben Hamida, Fabio Kon, Gustavo Ansaldi
Oliva, Carlos Eduardo Moreira Dos Santos, Jean-Pierre Lorré, Marco Autili,
Guglielmo De Angelis, Apostolos Zarras, Nikolaos Georgantas, Valérie Issarny,
and Antonia Bertolino, sketched technological solutions for future ultra large
XIV Future Internet Foundations
systems, addressing scalability and heterogeneity issues in order to leverage

Service-Oriented Architectures to support a wider range of services and users.
Thus, an architecture combining both a development and a runtime environment
is defined, as undertaken in the CHOReOS project.
Finally, the work by James Davey, Florian Mansmann, Jörn Kohlhammer,
and Daniel Keim on “Visual Analytics: Towards Intelligent Interactive Internet
and Security Solutions” presents an introduction to Visual Analytics and its rele-
vance to the Future Internet in particular the two facets, which are characterized
by a vast and growing amount of data: content and infrastructure. It shows that
emerging data visualization platforms for the web derive their value from the
relevance of the data that is analyzed with them. This paper argues that tar-
geted research in Visual Analytics can revolutionize the way in which humans
interact with content in the Future Internet.
Besides its potential for content, Visual Analytics can play an important role
in the network infrastructure of the Future Internet. Due to the amount of data
available from networking devices, the inherent complexity of the network and
the need to immediately react to failures or attacks, visual and computational
support for tasks in this domain can significantly improve infrastructure plan-
ning and testing, as well as network monitoring and security. Strengthening the
connection between Visual Analytics and the Future Internet would enable a
more secure, reliable and scalable infrastructure.
Future Internet Application Areas
Introduction
Stamatis Karnouskos
SAP Research
Petros Daras
Informatics and Telematics Institute, CERTH
Henning Müller
University of Applied Sciences Western
Switzerland, Sierre (HES-SO)
Man-Sze Li
IC Focus
Hans Schaffers
Aalto University
1 Introduction
Applications are software systems perceived and utilized by their intended users
to carry out a specific task. Applications are what users are actually using in their
working environments and their daily lives, hence applications are the medium
that enables them to interact with the rapidly advancing technologies. This im-
plies that we should take the users’ needs and aspirations as a point of departure
for developing and introducing advanced applications. Therefore, it is extremely
important to pay attention to the openness of the process of developing, testing
and validating applications and to the involvement of users in that process.
Applications evolve as they depend on the capabilities provided by several
real systems. For example, the end-user devices they run on as well as virtual
resources they utilize e.g. for mash-up applications are depending on the dis-
tributed services that provide the functionalities needed by these applications.
In the Future Internet (FI) era, the applications will enjoy both the advances we
have seen on the hardware e.g. running on mobile devices such as smartphones
with memory and CPU power that comparable to supercomputers a couple of
decades ago, as well as on the software side, where virtualization of the infrastruc-
ture and real-time communication and computation on data is possible. Taking
advantage of the rich information offered by various stakeholders as well as the
FI platform core facilities, the FI applications are expected to be seamlessly ad-
justing to the user’s needs and context, while in parallel hiding the complexity
of the underlying infrastructure and the interactions with the other services and
systems.
Some of the key Internet-based technologies underlying smart Future Internet
applications include cloud computing, real-world user interfaces of cyber-physical
systems and the semantic web. Cloud computing, a new way of delivering com-
XVI Future Internet Application Areas
puting resources, will have considerable impact as it is opening up new possibili-

ties in virtualization of physical spaces. The rapid advance of Internet of Things
technologies will enable us to sense the real-world and will empower a new class
of applications that are able to receive real-time information from the physical
surrounding and interact with it. For example, a new generation of location-
aware applications and services and, on the longer term new types of spatial
intelligence, will advance the end-user application capabilities while it will blend
easily in a global ecosystem of web applications, social media and crowdsourcing.
Finally, the semantic web is expected to facilitate the merging of data from dif-
ferent sources and presenting them in a meaningful way, thus bringing social
media based collaboration and collective intelligence to a higher level.
As the Future Internet will be a very complex system of systems, the ap-
plications will be the entry point for many users to interact with it and enjoy
its offerings. In such a complex and rapidly changing environment the applica-
tion developer will have to deal with multiple heterogeneous information sources
that will need to be integrated as well as an increasingly number of heteroge-
neous devices that will be used to interact with the user. To achieve this, we
will witness further increase in the trend to go beyond monolithic applications
towards composite ones that collaborate both with parts of the Future Internet
infrastructure as well as with other services and apps. This collaborative way of
interactions is expected to lead to emergent behaviours in the Future Internet
that at the end will better serve the end-users.
Several challenges need to be mastered in order to empower the visions
for highly sophisticated Future Internet Applications. The challenges are in-
creasingly multi-domain (ranging from technical to social, design, economics
etc.) while in parallel traditional issues such as security, trust, privacy, user-
friendliness and rapid development will still need to be present from day 1 and
not added as an aftermath. Complexity management, crowdsourcing, real-time
analytics, knowledge capturing and communication, simulation are only some
indicative aspects that will need to be investigated as they will impact the next
generation of applications. Nevertheless the Future Internet applications are in
the heart of emerging visions for a smarter world i.e. smart cities, smart energy,
smart health, smart enterprises, smart environment, smart transportation, lo-
gistics and mobility, smart manufacturing, smart agriculture and tourism. Their
existence has the challenging goal of enabling innovation by empowering the
Future Internet users.
2 Papers in the Section Applications

Various papers collected in this section demonstrate aspects of the scope and
width of advanced applications based on Future Internet technologies. The pa-
pers vary from offering the technology orientation of applications to demonstrat-
ing the importance of applications within various application contexts such as
service marketplaces and social networking.
More specifically, I-Search: A Unified Framework for Multimodal Search and
Retrieval focuses on novel approaches for multimodal search allowing for easy
Future Internet Application Areas XVII
retrieval of diverse media types simultaneously e.g. 3D objects, images, audio

and video. These technologies show a high potential value for enabling Internet-
based applications in various important sectors, which are characterized by an
overwhelming amount of content and could further serve for providing generic
enablers to FI-WARE in terms of accessing varied content.
Supporting content, context and user awareness in Future Internet applica-
tions presents the general idea of delivering complex services in a distributed
networking environment to end-users. The main feature of the proposed idea
is that the process of complex services delivery is aware of the content being
delivered, the context of the services delivery and that the delivered services are
personalized for each separate end-user.
Towards Trustworthy Marketplaces for Services and Apps in the Future In-
ternet presents the concept of trusted service marketplaces playing a key role for
the future Internet of Services. Such service marketplaces impose new demands
and requirements to trust and security and the paper proposes an approach and
vision to address these demands.
Semantically Enriched Services to Understand the Need of Entities describes
the evaluation of the proposed “Net-Ontology” aiming to improve network com-
munication with semantics. The paper addresses intermediate network layers
and contains an experimental evaluation and a promising comparison against
the current TCP/IP stack.
Using Future Internet Infrastructure and Smartphones for Mobility Trace
Acquisition and Social Interaction Monitoring focuses on the social network-
ing context of Future Internet applications, which is of high relevance to smart
city environments. The authors discuss a system for producing traces for a new
generation of human-centric applications, utilizing technologies such as Blue-
tooth and focusing on human interactions. Two deployments in human-centric
environments are described, one in an office environment and one in an exhibi-
tion/conference environment. The paper demonstrates the growing interaction
between technology development and user interaction,
3 Conclusions
The Future Internet will be information driven and rely on services to empower
the interactions among its stakeholders at multiple layers which will be facili-
tated via the applications. This calls for open information exchange and a new
generation of highly sophisticated applications customized to end-user needs.
Many of the papers in this volume are dealing with several central schemes of
the Internet: content of an increasingly unstructured nature such as images and
mixed media, needs to navigate this content using the user context such as lo-
cation and other sensors and trust in the information that is received and being
transmitted. All these application domains will change the way people interact
and the way that living spaces are being created, highlighting the impact of the
Future Internet on the lives of most of us.
Future Internet Application Areas: Smart Cities
Introduction
Srjdan Krco
Ericsson
Cities are complex, dynamic environments, catering for the needs of a large num-
ber of citizens and businesses (“users of city services”). The number of people liv-
ing in the cities globally is continuously increasing, we are witnessing emergence
of mega cities and the need for a sustainable development of such environments
is more than evident.
The smart cities concept is not new. A number of cities around the world
are using “smart” designation aiming to show that they have already done some-
thing in that regard or are planning to do so (this ranges from deploying optical
infrastructure across a city to introduction of e-government services or some
other, mainly ICT based, improvements making the city services more efficient
or quality of life of the citizens better). With the recent technological advances
in the domains of Internet of Things, M2M, big data, and visual analytics and
leveraging the existing extensively deployed ICT infrastructure, the smart cities
has attracted a lot of interest over the last few years.
Combining these technologies has made it possible to improve a range of the
city services, from the public transport domain and traffic management to the
utility services like water and waste management and public security and safety.
All these services are intrinsically connected and interweaved. Therefore, for a
city to develop in a sustainable and organized manner it is crucial to coordinate
such developments and make it possible for smart services to leverage each others’
functionality. The city governments will have a crucial role in these endeavors,
from the overall city planning perspective as well as creation of the regulation
and legislation framework for smart city service developers and providers.
The content of this area includes three chapters covering smart cities from
three different perspectives: social, legislation and safety.
The “Towards a Narrative-Aware Design Framework For Smart Urban En-
vironments” chapter is focusing on smart cities from both technical and social
perspectives. The chapter describes a new narrative-aware design framework for
the smart cities which combines quantitative sensor-generated data (Internet of
Things installations) as well as qualitative human generated data (human story-
telling) through participatory web platforms, in an always-on networked world.
Three levels are identified in the framework: “data and stories”, “analysis and
processing” and “services and applications”. Examples of narrative-aware urban
applications based on the design framework are given and analyzed.
The “Urban Planning and Smart Cities: Interrelations and Reciprocities”
chapter analyses the smart city’s contribution in the overall urban planning and
XX Future Internet Application Areas: Smart Cities
vice versa. It highlights and measures smart city and urban planning interrela-
tion and identifies the meeting points among them. The chapter starts with the
urban planning principles based on the European Regional Cohesion Policy, and
then identifies the key smart city attributes and characteristics. Finally, analysis
of the way these domains influence each other and impact development of each
domain is given.
The “The Safety transformation in the Future Internet domain” chapter is
dealing with the public safety as one of the major concerns for governments and
policy makers in smart cities. The chapter presents an introduction to Inter-
net of things, Intelligent Video Analytics and Data Mining Intelligence as three
fundamental pillars of the Future Internet infrastructure in the public safety
domain.
Future Internet Infrastructures
Introduction
Federico Alvarez
Universidad Politécnica de Madrid, Madrid, Spain
Elio Salvadori
CREATE-NET, Trento, Italy
Anastasios Gavras
Eurescom, Heidelberg, Germany
One of the most important aspects of the Future Internet is to leverage existing
investments in advanced infrastructures for testing and experimentation of novel
Future Internet technologies and speed up their introduction into the market.
A large number of advanced infrastructures are available in regions such as
Europe ranging from national or European level to many regional or city-level
initiatives promoting innovative FI concepts such as smart-cities, smart-grids
and e-health.
Europe has the potential to deliver massive capacity for Future Internet de-
velopments by leveraging the abundance of advanced infrastructures (current in-
frastructure, future infrastructure, pilot experiments, testbeds, and experimental
facilities) but the fragmentation and lack of interoperability and understanding
of capacities hinder those development at large scale.
The Future Internet Research and Experimentation (FIRE) initiative
(www.ict-fire.eu) in Framework Programme 7 created a research environment
for investigating and experimentally validating revolutionary ideas towards new
paradigms for Future Internet architecture by bridging multi-disciplinary long-
term research and experimentally driven large scale validation. FIRE invested
significant effort in familiarising the ICT research community with the method-
ology of experimental driven research as a necessary research tool in the ICT
related science disciplines.
In some cases it is difficult to define what is an ICT infrastructure. The
definition of infrastructures done in the project INFINITY (www.fi-infinity-eu)
is the following:
An infrastructure is an structured and organised collection of physical and/or
logical elements offering an ICT platform with the functionality to facilitate large
scale experimentation and testing for Future Internet projects and applications
and service developments.
Such a platform may consist of ICT-based services which could be generic
or more specific to a given domain (e.g. energy, transport, health, environment,
tourism, health. . . ).”
In consequence, ICT based infrastructures are one of the chapters which this
book is addressing.
XXII Future Internet Infrastructures
Following the review and selection process run for this FIA book, four chap-
ters were chosen. The following is a summary of main results of the “Future
Internet Foundations” section of this FIA book.
The paper on “FSToolkit: Adopting Software Engineering practices for
enabling definitions of federated resource infrastructures” by Christos Tranoris,
and Spyros Denazis describes the present the Federation Scenario Toolkit
(FSToolkit) that enables the definition of resource request scenarios, agnostic
in term of providers. This work adopts Software Engineering practices consider-
ing the concepts of modeling and meta-modeling to define a resource broker and
to specify scenarios by applying the Domain Specific Modeling (DSM) paradigm.
FSToolkit is developed for experimentally driven research for validating through
testing-scenarios new architectures and systems at scale and under realistic en-
vironments by enabling federation of resources.
The work by Leonidas Lymberopoulos, Mary Grammatikou, Martin Potts,
Paola Grosso, Attila Fekete, Bartosz Belter, Mauro Campanella and Vasilis
Maglaris“NOVI Tools and Algorithms for Federating Virtualized Infrastructures”
addresses the efficient approaches to compose virtualized e-Infrastructures to-
wards a holistic Future Internet (FI) cloud service and aspires to develop and
validate methods, information systems and algorithms that will provide users
with isolated slices, baskets of resources and services drawn from federated in-
frastructures.
The paper “Next Generation Flexible and Cognitive Heterogeneous Optical
Networks Supporting the evolution to the Future Internet” by Ioannis Tomkos,
Marianna Angelou, Ramón J. Durán Barroso,Ignacio de Miguel, Rubén Lorenzo,
Domenico Siracusa, Elio Salvadori, Andrzej Tymecki, Yabin Ye and Idelfonso
Tafur Monroy describes the new research directions in optical networking to
further advance the capabilities of the Future Internet. They highlight the latest
activities of the optical networking community and propose concepts of flexible
and cognitive optical networks including their key expected benefits..
The work by Marc Pallot, Brigitte Trousse, Bernard Senach “A Tentative
Design of a Future Internet Networking Domain Landscape” presents a tentative
FI domain landscape populated by Internet computing and networking research
areas where still open questions such as visualzsing the conceptual evolution
and articulating the various FI networking and computing research areas and
identifying appropriate concepts populating such a FI domain landscape are
developed.
Table of Contents
Invited Introduction Paper

Introduction: FIA Research Roadmap, Priorities for Future Internet
Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Nick Wainwright and Nick Papanikolaou
Foundations
A Tussle Analysis for Information-Centric Networking Architectures . . . . 6
Alexandros Kostopoulos, Ioanna Papafili, Costas Kalogiros,
Tapio Levä, Nan Zhang, and Dirk Trossen
A Systematic Approach for Closing the Research to Standardization

Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Bernard Sales, Emmanuel Darmois, Dimitri Papadimitriou, and
Didier Bourse
SOCIETIES: Where Pervasive Meets Social . . . . . . . . . . . . . . . . . . . . . . . . . 30

Kevin Doolin, Ioanna Roussaki, Mark Roddy, Nikos Kalatzis,
Elizabeth Papadopoulou, Nick Taylor, Nicolas Liampotis,
David McKitterick, Edel Jennings, and Pavlos Kosmides
Cross-Disciplinary Lessons for the Future Internet . . . . . . . . . . . . . . . . . . . . 42

Anne-Marie Oostveen, Isis Hjorth, Brian Pickering,
Michael Boniface, Eric T. Meyer, Cristobal Cobo, and
Ralph Schroeder
Design Principles for the Future Internet Architecture . . . . . . . . . . . . . . . . 55

Dimitri Papadimitriou, Theodore Zahariadis, Pedro Martinez-Julia,
Ioanna Papafili, Vito Morreale, Francesco Torelli,
Bernard Sales, and Piet Demeester
From Internet Architecture Research to Standards . . . . . . . . . . . . . . . . . . . 68

Dimitri Papadimitriou, Bernard Sales, Piet Demeester, and
Theodore Zahariadis
An Integrated Development and Runtime Environment for the Future

Internet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Amira Ben Hamida, Fabio Kon, Gustavo Ansaldi Oliva,
Carlos Eduardo Moreira Dos Santos, Jean-Pierre Lorré,
Marco Autili, Guglielmo De Angelis, Apostolos Zarras,
Nikolaos Georgantas, Valérie Issarny, and Antonia Bertolino
XXIV Table of Contents
Visual Analytics: Towards Intelligent Interactive Internet and Security

Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
James Davey, Florian Mansmann, Jörn Kohlhammer, and
Daniel Keim
Applications
Towards a Trustworthy Service Marketplace for the Future Internet . . . . 105
Francesco Di Cerbo, Michele Bezzi, Samuel Paul Kaluvuri,
Antonino Sabetta, Slim Trabelsi, and Volkmar Lotz
Using Future Internet Infrastructure and Smartphones for Mobility

Trace Acquisition and Social Interactions Monitoring . . . . . . . . . . . . . . . . . 117
Athanasios Antoniou, Evangelos Theodoridis,
Ioannis Chatzigiannakis, and Georgios Mylonas
I-SEARCH: A Unified Framework for Multimodal Search and

Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Apostolos Axenopoulos, Petros Daras, Sotiris Malassiotis,
Vincenzo Croce, Marilena Lazzaro, Jonas Etzold, Paul Grimm,
Alberto Massari, Antonio Camurri, Thomas Steiner, and
Dimitrios Tzovaras
Semantically Enriched Services to Understand the Need of Entities . . . . . 142

Flávio de Oliveira Silva, Alex Dias, Caio César Ferreira,
Eduardo De Souza Santos, Fabı́ola Souza Fernandes Pereira,
Isabelle Cecı́lia de Andrade, João Henrique de Souza Pereira,
Lásaro Jonas Camargos, Luiz Cláudio Theodoro,
Maurı́cio Amaral Gonçalves, Rafael Pasquini,
Augusto José Venâncio Neto, Pedro Frosi Rosa, and
Sergio Takeo Kofuji
Supporting Content, Context and User Awareness in Future Internet

Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Pawel Światek,
Krzysztof Juszczyszyn, Krzysztof Brzostowski,
Jaroslaw Drapala, and Adam Grzech
Smart Cities
Towards a Narrative-Aware Design Framework for Smart Urban
Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Lara Srivastava and Athena Vakali
Urban Planning and Smart Cities: Interrelations and Reciprocities . . . . . 178

Leonidas G. Anthopoulos and Athena Vakali
Table of Contents XXV
The Safety Transformation in the Future Internet Domain . . . . . . . . . . . . 190

Roberto Gimenez, Diego Fuentes, Emilio Martin, Diego Gimenez,
Judith Pertejo, Sofia Tsekeridou, Roberto Gavazzi,
Mario Carabaño, and Sofia Virgos
Infrastructures
FSToolkit: Adopting Software Engineering Practices for Enabling
Definitions of Federated Resource Infrastructures . . . . . . . . . . . . . . . . . . . . . 201
Christos Tranoris and Spyros Denazis
NOVI Tools and Algorithms for Federating Virtualized

Infrastructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Leonidas Lymberopoulos, Mary Grammatikou, Martin Potts,
Paola Grosso, Attila Fekete, Bartosz Belter,
Mauro Campanella, and Vasilis Maglaris
Next Generation Flexible and Cognitive Heterogeneous Optical

Networks: Supporting the Evolution to the Future Internet . . . . . . . . . . . . 225
Ioannis Tomkos, Marianna Angelou, Ramón J. Durán Barroso,
Ignacio de Miguel, Rubén M. Lorenzo Toledo, Domenico Siracusa,
Elio Salvadori, Andrzej Tymecki, Yabin Ye, and
Idelfonso Tafur Monroy
A Tentative Design of a Future Internet Networking Domain

Landscape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Marc Pallot, Brigitte Trousse, and Bernard Senach
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Introduction: The FIA Research Roadmap,
Priorities for Future Internet Research
Nick Wainwright and Nick Papanikolaou
Cloud and Security Lab, HP Labs

{nick.wainwright,nick.papanikolaou}@hp.com
Abstract. We describe the key findings of the Future Internet Assem-

bly Research Roadmap for Framework Programme 8, which captures the
ideas and contributions of the FIA community on the important research
topics that should be addressed in future funding programmes. The find-
ings of the roadmap have been produced through an open consultation of
research projects who participate in FIA. It is primarily concerned with
identifying research that can be carried out in the second half of this
decade and which will have an impact in 2020 and beyond. By ‘impact’
we mean will result in products, services, systems, capabilities, that come
to market and are available and deployed in that timeframe.
Introduction. We present here a vision for the Future Internet and its impact
on individuals, businesses and society as a whole; the vision presented is based
on an extended consultation carried out by the authors within the European
Future Internet research community, as part of the work of the Future Internet
Assembly (FIA).
The purpose of the consultation was to identify key challenges and research
priorities for the Future Internet, particularly from the standpoint of current
European research projects (in Framework Programme 7). The output of the
consultation is documented in the form of a visionary research roadmap1 .
In order to elicit inputs from members of the European Future Internet re-
search community, we have had to actively participate in this community our-
selves; the vehicle for doing so has been the EU Framework Programme 7 research
project EFFECTSPLUS2 , which carries out workshops and clustering activities
for European projects, particularly in the area of ICT trust and security. As part
of this Support Action we participate in and run aspects of the Future Internet
Assembly (FIA).
The vision we present is intended to inform future research funding pro-
grammes, including the European Commission’s ”Horizon 2020” framework pro-
gramme. We have validated the results of our initial consultation, and the
associated vision, with a significant number of researchers in the FI commu-
nity, and in this paper we also present additional insights gained during this
1
This is available online at http://fisa.future-internet.eu/index.php/
FIA_Research_Roadmap
2
See www.effectsplus.eu
F. Álvarez et al. (Eds.): FIA 2012, LNCS 7281, pp. 1–5, 2012.

c The Author(s). This article is published with open access at SpringerLink.com
2 N. Wainwright and N. Papanikolaou
process of validation. Overall, we observe that there are several important areas
of innovation for Future Internet research and that these need to be developed
and supported by researchers and policymakers both within and outside Europe.
Research Priorities. In presenting the research needs in this roadmap we have

brought out the ‘horizontal challenges’ — those that underpin most, if not all, of
the applications and use of Future Internet. In developing this roadmap we have
had specific inputs in the areas of smart cities, the future of enterprise, digital
media, and inclusion. We had little input in areas such as healthcare, energy
management, transportation. We don’t think that is too much a limitation —
our aim has been to bring out the broad and horizontal research themes that will
persist, even grow in scope and opportunity and applicability as we look past
the middle of the decade, and particularly those area that address fundamental
capabilities, and needs of a networked information society.
In our research priorities for the Future Internet we see three themes that are
foundational and enabling and which support us in building the Future Internet.
These are:
1. Going Beyond Converged Infrastructure: the Internet infrastructure

beyond 2020 brings new capabilities and capacities;
2. The Rise of Networked Data: learning to exploit the Internet’s natural
resource;
3. Achieving Real Internet Security: maintaining the security of the In-
ternet and its users online.
Secondly, looking forward towards the research that will transform what we do
and how we do it and which are fundamentally integrative, they exploit and use
a wide range of networked technologies towards a diverse set of objectives, we
see three priorities that support us in using the Future Internet. These are:
4. Networked Interaction: people interacting with each other, with infor-

mation, and with cyber-physical worlds;
5. Augmented Worlds: from an Internet of things to an Internet doing
things;
6. Internet-style Innovation: the Internet as an innovation ecosystem, sup-
ported by architecture, policy, and invention.
Going Beyond Converged Infrastructure. Internet infrastructures — net-

works, virtualised computing, storage systems, undergoing a period of intense
convergence, the boundaries between service platforms and the infrastructure
services layer are becoming blurred, and at the same time at the edge of the
networks smart mobile devices are becoming pervasive and more capable at the
edges of networks.
Future research must look beyond converged infrastructure to the addressing
the challenges of meeting the vastly increased demand for bandwidth and services
and connectivity and new and different applications and services:
The FIA Research Roadmap, Priorities for Future Internet Research 3
– Polymorphic networks — combining different networks to meet the for ca-

pacity and needs of new media, applications, services, infrastructures and
networked ‘things’
– Expanding the cloud to the edges of the network and beyond, providing
the execution environments for new FI applications and services. Real time
capabilities will be vital for these new services, data and event processing,
interaction processing all demand real time responses, and variability in de-
mand will provide real challenges for the services providers
– Looking beyond smart devices towards smart edge-systems as the execution
environments at the edges of the network that link the physical and the
cyber world.
The Rise of Networked Data. Networked data on a massive scale is the

powerhouse of the Internet today and the growth trend looks set to continue as
new services and applications are developed, a greater part of the economy and
public sector relies on the Internet, and citizens spend increasing parts of their
lives online.
Networked data is a horizontal capability of the Future Internet. The organisa-
tion, exploitation, and governance of the huge amount of data and information
in the Internet will create ongoing opportunities and challenges for the next
decade. There are underlying tensions between rights of citizens, businesses, and
the state over data; the opportunities opened up by integrating data from mul-
tiple sources; the need develop new data models to make sense of the myriad
of applications and sources (e.g. the 3D world for augmenting spaces, of tacit
knowledge in knowledge supply chains), of records of usage, of surveillance data
gathered, the list continues to grow.
Achieving Real Internet Security. A decade away we will be conducting

much more of our lives and economy online than we do now. The scale of In-
ternet use continues to increase relentlessly, and our reliance on the Internet
continues to increase. As we do so the potential opportunity, rewards, and im-
pact of cybercrime becomes even more significant. The scale of threats, potential
for conflict between individuals, organisations, and states online should not be
underestimated. Reliance on Internet for operation of our critical infrastruc-
tures means that cyber defence is an even more vital aspect of state security;
particular concerns here are cyber attacks by unfriendly states, as well as cyber
interventions by activists. Cyber attacks will become industrialised, and we need
to guard against organised cybercrime as well as targeted attacks on individuals
or particular systems.
Securing the Internet as a socio-technical system is a high priority and con-
tinued separate attention is necessary. Internet security can become part of Eu-
rope’s new defence businesses, and a secure Internet is necessary condition for
economic competitiveness. It is imperative to make the Internet safe and secure,
so it is a positive experience for all users, independently of their background or
education. At the same time, we need to make security controls transparent and
4 N. Wainwright and N. Papanikolaou
unobtrusive, so that they do not hinder openness and availability of network

resources.
Networked Interaction. Through the Internet we interact with each other,

with the physical world, and with the digital world, and indeed, in the future
the distinctions my blur even further. Social networks are not the last word
on social interaction, webcams and video conferencing are not the last word on
collaboration, games and IPTV are not the last word on entertainment. New
interfaces and modalities will create opportunities for richer interaction and for
addressing our work, life and emotional needs. New ways of interacting with
complex data provide ways to understand complex situation. New interactions
with the digital world will provide new media experiences that look beyond 3d.
Future networked interaction will not be delivered through one device, in a sit-
back, sit-up, or handheld interaction mode, but through collections of devices
brought together as smart edge systems, and ideas of ownership, situatedness,
virtualisation will create interaction experiences that are effective, engaging, and
empowering.
The research theme of interaction, supported by rich interfaces, displays, hap-
tics, and other yet to be developed approaches makes possible to address some of
our real concrete needs too — for carbon reduction for example through remote
collaboration which can take people off roads, support knowledge business net-
works, or create valuable social links. Some of the biggest barriers to delivering
Internet benefits to excluded groups in Internet are the interfaces. This is just
one example, and looking forward the future of networked interaction has real
potential to create value given that we are reaching the point where demand
and capability come together to make new and valuable networked interactions
possible.
Augmented Worlds. The vision of networked services, systems, and devices

supporting us in our work and social lives, or in business to control and manage
processes and operations has been with us for some time. Hitherto, the Frame-
work Programme Seven has explored Internet of services, reflecting the shift of
our economy to a service economy, Internet of things, reflecting the opportunity
to measure and manage the physical word using networked systems and these
capabilities are beginning to be available.
As we look to the start of the next decade and beyond we can begin explore
how we can harness the power of the Internet to augment lives, work, business
and spaces in ways that add value. By ‘augmentation’ we mean ‘increasing in in-
tensity’ the activities we are doing or the things we need done for us, addressing
what we do in our jobs and daily lives, addressing needs of groups and com-
munities, of industry, construction, maintenance, engineering, manufacturing,
transport with information, decisions support, risk analysis, options, delivered
through interactions and interfaces that are intuitive and unintrusive. What is
currently described as augmented reality has potential to develop into what is
fundamentally an integrative, systems, applied approach to addressing prob-
lems of industry, people, society and developing techniques and frameworks that
The FIA Research Roadmap, Priorities for Future Internet Research 5
harness the scale of the network and networked data onto individual actions,
tasks, and activities, transforming what we do and how we do it.
Internet-style Innovation. It is clear that the Internet has been an incredible

force for innovation over the past three decades at least. To have created such
a platform for innovation, value creation, and benefits to society must rank as
one of the outstanding achievements since the industrial revolution. It is our aim
that this innovation and value creation should continue, and that Europe should
play a big part in it. Every one of the research communities consulted in this
programme stressed innovation — both within their field of work and enabling
innovative benefits as a consequence of it.
Whether we are discussing topics such future enterprise, cities, or experiences,
ideas abound on the kinds of approaches that enable innovative value creation
to take off. The network effect, scale, openness, experimentation, software, and
pilots, and services, SME and start up participation, application, and real users.
This is the ‘Internet-style’ innovation we aim for. Innovation that happens at
the edges of the network, that is stimulated by linkages between sectors, that
involves people, where people, communities, business, even public sector, are
‘empowered’ to take control of opportunities to innovate. Ideas such as making
cities into experimental services environments, creating platforms, integrating
across industries and sectors, releasing and exploiting data, are enabling factors
— they set the conditions for unlocking value and if carried out ‘Internet-style’
they set the conditions for innovation and provide an environment for new ap-
plications and services need to be instantiated, built, used, and grow.
Conclusions. We have highlighted the themes above because they need to be

directly addressed and they deal with different aspects of the Future Internet.
There are also a number of approaches that are essential to the success of Future
Internet initiatives, where we discuss approached to Future Internet research.
As FI develops support for a wide range of stakeholders seeking to develop,
provision, or use a range of networked components and concepts there needs
to be an architectural framework that provides ongoing guidance, specification
and rules of how systems should behave, how everything fits together, how net-
worked elements communicate, and how elements are (dynamically) structured
into larger interoperating entities. The architecture needs to present the whole
picture, to relate the relevant elements in the picture, and to maintain its own
forward plan or roadmap as a consistent part of the overall FIA roadmap. Such
architectural coordination has typically been provided by groups such as IETF.
As we look forward to research in Future Internet we emphasise the need for the
research to take an architectural approach at all levels from infrastructure to
services and applications, and to participate in, and where appropriate develop
new forums for providing and developing that architectural coordination.
Open Access. This article is distributed under the terms of the Creative Commons
Attribution Noncommercial License which permits any noncommercial use, distribu-
tion, and reproduction in any medium, provided the original author(s) and source are
credited.
A Tussle Analysis
for Information-Centric Networking Architectures
Alexandros Kostopoulos1, Ioanna Papafili1, Costas Kalogiros1, Tapio Levä2,

Nan Zhang2, and Dirk Trossen3
1
Athens University of Economics and Business, Department of Informatics, Athens, Greece
2
Aalto University, Department of Communications and Networking, Espoo, Finland
3
University of Cambridge, Computer Laboratory, Cambridge, UK
{alexkosto,iopapafi,ckalog}@aueb.gr,
{tapio.leva,nan.zhang}@aalto.fi, [email protected]
Abstract. Current Future Internet (FI) research brings out the trend of designing
information-oriented networks, in contrast to the current host-centric Internet.
Information-centric Networking (ICN) focuses on finding and transmitting
information to end-users, instead of connecting end hosts that exchange data. The
key concepts of ICN are expected to have significant impact on the FI, and to
create new challenges for all associated stakeholders. In order to investigate the
motives as well as the arising conflicts between the stakeholders, we apply a
tussle analysis methodology in a content delivery scenario incorporating socio-
economic principles. Our analysis highlights the interests of the various
stakeholders and the issues that should be taken into account by designers when
deploying new content delivery schemes under the ICN paradigm.
Keywords: information-centric networking, content delivery, future internet

architecture, tussles, incentives, socio-economics, value network.
1 Introduction
Over the recent years, an increasing number of users gain access to the Internet via
numerous devices equipped with multiple interfaces, capable of running different
types of applications, and generating huge data traffic volumes, mostly for content.
Traffic stemming out of these activities implies increased cost for the Internet Service
Providers (ISPs) due to the congestion in their networks and the generated transit
costs, as well as unsatisfactory Quality of Service (QoS) for some end-users.
This exponential growth of content traffic has been initially addressed by peer-to-
peer applications, or Content Distribution Networks (CDNs). CDNs consist of
distributed data centers where replicas of content are cached in order to improve users’
access to the content (i.e., by increasing access bandwidth and redundancy, and
reducing access latency). These CDNs practically formulate overlay networks [1]
performing their own traffic optimization and making content routing decisions using
incomplete information about customer’s location and demand for content, as well as
utilization of networks and available content sources. Similarly ISPs perform individual
traffic optimization using proprietary, non-native and usually non-scalable solutions for
F. Álvarez et al. (Eds.): FIA 2012, LNCS 7281, pp. 6–17, 2012.
© The Author(s). This article is published with open access at SpringerLink.com
A Tussle Analysis for Information-Centric Networking Architectures 7
traffic monitoring and shaping (e.g. Deep Packet Inspection (DPI) boxes for peer-to-
peer traffic) and have no incentive to reveal information about their network to CDNs.
This information asymmetry often leads to a suboptimal system operation.
Information-centric Networking (ICN) postulates a fundamental paradigm shift
away from a host-centric model towards an information-centric one. ICN focuses on
information item discovery and transmission and not on the connection of end-points
that exchange data. Thus, ICN has the potential to address efficiently the
aforementioned information asymmetry problem by including traffic management,
content replication and name resolution as inherent capabilities of the network.
What remains the same is that the Internet is a platform composed of multiple
technologies and an environment where multiple stakeholders interact; thus, the
Internet is interesting from both the technological and the socio-economic viewpoint.
Socio-economic analysis comprises a necessary tool for understanding system
requirements and designing a flexible and successful FI architecture.
A first attempt to investigate socio-economic aspects of FI in a systematic manner
was performed by Clark et al. [2]. They introduced the ‘Design for Tussle’ principle,
where the term ‘tussle’ is described as an ‘ongoing contention among parties with
conflicting interests’. It is obvious that the need for designing a tussle-aware FI has
emerged to enhance deployment, stability and interoperability of new solutions.
Although there are plenty of counter-examples of adopted protocols/architectures that
do not follow the Design for Tussle principle, tussle-aware protocols and architectures
are expected to have better chances for adoption/success in the long-term [3].
The need for understanding the socio-economic environment, the control exerted
on the design, as well as the tussles arising therein has been also highlighted in [4].
The purpose of this work is to explore and analyze the tussles that may arise in ICN,
as well as to consider the roles of different stakeholders; below, we present a tussle
analysis methodology which extends the methodology originally developed within the
SESERV project [5], and apply it in the content delivery scenario. We focus on the
tussle spaces of name resolution, content delivery and caching.
This paper is organized as follows: In Section 2, we present our methodology for
identifying tussles among different stakeholders. Then, Section 3 provides an
overview of representative information-centric networking architectures developed in
the PURSUIT [6] and SAIL [7] research projects. In Section 4, we focus on a use case
for content delivery; we identify the involved stakeholders and major functionalities
and roles that they can take, and then investigate the potential tussles among the
stakeholders. Finally, in Section 5, we conclude our remarks.
2 A Methodology for Tussle Analysis

This section provides a generic guide for better understanding the impact of a
technology on the stakeholders’ strategies, as well as on how other technologies might
be used and deployed. Below, we extend the methodology presented in [8] and
combine it with the Value Network Configuration (VNC) method introduced by
Casey et al. [9]. The tussle analysis methodology consists of the following steps:
1. Identify all primary stakeholder roles and their characteristics for the functionality
under investigation.
8 A. Kostopoulos et al.
2. Identify tussles among identified stakeholders.

3. For each tussle:
(a) Translate knowledge into models by assessing the mid-term and long-term
impact to each stakeholder;
(b) Identify potential ways for stakeholders to circumvent negative impacts, and
the resulting spill-overs.
4. For each circumventing technique, apply steps 1-4 again.
The involved stakeholders usually express their interests by making choices that will
affect the technology by deciding which technologies will be introduced, how these
will be dimensioned, configured, and finally, used. All these collective decisions will
eventually determine how technology components will operate and produce outputs
that are valuable for these stakeholders. Technology outputs are assessed by each
stakeholder individually and can affect real-world interactions (e.g. payments, price
competition, price regulation and collaboration) or trigger new technology decisions.
Such interactions allow the Internet to evolve and act as a living organism (Fig. 1).
Fig. 1. The Socio-Economic layer and Technology layer of the Internet ecosystem
Several techniques or methods can be used to perform each of the aforementioned

steps. In this paper, we show how the VNC method [9] can be incorporated in the
tussle analysis. What makes the VNC method a particularly useful tool for tussle
analysis is the separation of the stakeholders (or actors as Casey et al. call them) from
the functional roles the actors can take, thus allowing us to analyze multiple role
combinations instead of limiting to a single value network.
Identifying functional roles - defined in [9] as a set of activities and technical
components, the responsibility of which is not divided between separate actors in a
particular scenario- is central to the VNC method. Because roles hold economic and
strategic value, the actors fight for their control. The tussles emerge when there is a
conflict of interest between the actor controlling the role and the other actors affected
by it. Depending on which actor controls a role, the tussle outcomes and the
circumventing techniques vary, which further motivates the usage of the VNC
method.
The VNC method emphasizes technologies’ role in defining the possible value
networks by identifying also the technical components and technical interfaces
between them. By doing this, the method improves our understanding of the
relationship between the technical architecture (a set of technical components linked
to each other with technical interfaces, such as protocols) and the value network
configuration (role division and related business interfaces among actors). This is
important in analyzing whether the technology is designed for tussle [2], i.e., if the
technical design allows variation in value networks. Fig. 2 presents the notation
presented in [9] that can be used to visualize the roles and VNC.
Fig. 2. Notation of VNC methodology
After identifying the involved stakeholders as well as the tussles among them, the
next step is to translate knowledge into models and provide quantitative analysis. In
[10] a toolkit is suggested that uses mind-mapping techniques and system dynamics to
model the tussles. System Dynamics (SD) [11] is a useful tool to evaluate dynamic
interactions between multiple stakeholders, by simulating the possible outcomes (e.g.,
how technology diffuses) when multiple stakeholders interact. The main focus is on
the assessment of outcomes and their evolution over time, since possible reactions can
be modeled. After having captured the causality models, relevant socio-economic
scenarios may be formulated to investigate the potential consequences in the Internet
market. We do not conduct SD analysis in this paper due to space constraints.
3 Overview of ICN Architectures

Diverse research projects, such as PURSUIT [6], SAIL [7] and NDN [12] are
emphasizing the need to move towards an ICN architecture. In this section we briefly
present an architecture overview of ICN in order to provide the necessary
background. We focus on the Publish/Subscribe (pub/sub) model adopted by
PURSUIT and the Network of Information (NetInf) introduced by SAIL.
3.1 Publish/Subscribe
In the PURSUIT pub/sub paradigm, information is organized in scopes. A scope is a
way of grouping related information items together. A dedicated matching process
ensures that data exchange occurs only when a match in information item (e.g., a
video file) and scope (e.g., a YouTube channel) has been made. Each packet contains
the necessary meta-data for travelling within the network. Fig. 3 presents a high level
picture of the main architectural components of the pub/sub architecture.
Fig. 3. A Publish/Subscribe architecture for ICN [13]
At the application level, the pub/sub components implement applications based on

basic ICN services, enabling publications and subscriptions towards information
items within particular scopes.
At the network level, the architecture itself consists of three main functions:
rendezvous, topology and forwarding. The rendezvous function implements the
matching between publishers and subscribers of information based on several criteria.
Moreover, the rendezvous service provides additional functionalities to implement
policies associated with the matching, such as access control. When a publication is
matched with one or more subscriptions, an inter-domain forwarding graph is created
in negotiation with the inter-domain topology formation (ITF) function. After
constructing inter-domain paths between the forwarding networks to which
publisher(s) and subscriber(s) are attached, intra-domain paths need to be constructed.
This is done in collaboration with the AS-internal topology management (TM)
function, which instructs its local forwarding nodes (FN) to establish paths to local
publishers / subscribers or to serve as transfer links between ASes.
3.2 Network of Information

The SAIL Network of Information (NetInf) aims at three architectural objectives: i)
unique naming regardless of the Named Data Object’s (NDO’s) location and without
a hierarchical naming structure; ii) receiver-oriented NDO delivery; and iii) a multi-
technology and multi-domain approach, where any underlying technology and
network can be leveraged [14]. The NetInf network consists of Name Resolution
System (NRS) nodes and NetInf router (NR) nodes, which are illustrated in Fig. 4.
NetInf supports both name-based routing and name resolution. Name resolution is
enabling scalable and global communication: NDOs are published into the network
and registered by the NRS. Specifically, the NRS is used to register the network
locators of NDO copies in the underlying network, which can potentially provide
packet-level routing and forwarding functionalities. The NDO request can be resolved
by the NRS into a set of network locators, which are used to retrieve a copy of the
NDO from the optimum source based on a pre-defined criterion. At least one global
NRS must exist in the NetInf network, but also intra-domain NRS’ are possible.
The NetInf router node accepts NetInf names as input and decides how to route the
request so that eventually a NDO is returned to the previous-hop NetInf node. This
routing decision could be either towards a NRS or directly towards the NDO source, the
latter of which represents the name-based routing scenario. In addition, NetInf cache
servers for content replication can be placed both in the NR nodes and the NRS nodes.
Fig. 4. NetInf high level architecture
Fig. 4 also shows the high level content retrieval process in NetInf. First, (1) a
NDO owner publishes the NDO into the network by adding it to the NRS registry.
When a (2) request for a NDO occurs, the NetInf router can either (3a) forward the
request to a NRS for (3b) the set of locators or it can (4) directly forward the request
to the NDO source, depending on whether the NetInf router knows where the NDO is.
Finally, (5) the NDO is returned to the requester via the same route as the request and
the NDO can be cached on every node that it passes.
4 Tussles in Information-Centric Networking

In this section, we focus on the content delivery use-case in a generic ICN
architecture and apply our combined tussle analysis and VNC methodologies to it. We
first look into the intra-domain scenario and then build incrementally on the inter-
domain scenario. As the first step of our methodology, we identify here major
functionalities, group them into roles and list the stakeholders that can take up these
roles. Then, in the second step, we perform tussle analysis on a per functionality view.
4.1 The Content Delivery Use-Case

As illustrated in Fig. 5, we consider two Access Network Providers (ANPs) that
employ ICN to offer content delivery services to their customers. The two ANPs are
connected through transit links to an Inter-Connectivity Provider (ICP). Both ANPs

employing ICN have deployed their own networks of Caches. Within the ANPs
premises, local NRSs are also provided, which are connected to a global NRS service.
The NRSs could be controlled by either the respective network infrastructure provider
(ANP or interconnectivity provider) itself, or by a third-party. Potential subscribers of
an information item exist in both ANPs; however, only a single publisher (P1) of that
specific content exists initially, in ANP1.
Fig. 5. Content delivery in ICN architecture
Intra-domain Scenario. We assume that P1 in ANP1 publishes an information item to

his local NRS, and the local NRS advertises the publication to the global NRS. Then,
S1 in ANP1 sends a subscription for an information item to the local NRS of its ANP.
The local NRS identifies that the requested information item is published within the
ANP and matches P1 with S1. If more subscriptions for the same information item
occur, the ANP may also decide to cache the content to another location in order to
achieve load balancing and to provide higher QoS to its customers (subscribers).
Inter-domain Scenario. Let us now assume that S2 in ANP2 also subscribes to his
local NRS for the same information item. Since, the information item is not published
within ANP1, the local NRS informs the global NRS about this subscription. The
global NRS, who is aware of P1, matches P1 with S2. ANP2 may cache the information
item in his caching elements, in order to serve potential new subscribers.
4.2 Functionalities, Roles, Stakeholders

Based on the aforementioned use-case, we identify the key functionalities and map
them to five key roles (Table 1). There are multiple stakeholders in position to control
these roles, which would lead to different outcomes. Here, we focus on the role
allocation visualized in Fig. 6, since it is a representative case to take place in ICN. In
our setup, the content access management (i.e. AAA) role can be taken by either the
Content Provider (CP) or the ANP, the name resolution is taken by either the ANP or
a third-party provider (i.e. a Rendezvous Network (RENE) provider in [6]), whereas
the other four roles are assigned to the ANP. The chosen role allocation differs from
the typical situation in the market today where other stakeholders, such as CDN
providers or CPs, control the name resolution, caches and content network.
Table 1. Key roles and functionalities in ICN content delivery
Role Functionalities
Name Resolution Content directory control, names to locations resolution,
rendezvous, matching, applying policies
Content access management AAA (Authentication, Authorization, Accounting)
Cache management Cache servers control, content selection for being cached,
cache updating
Cache location ownership Cache locations control
Content network management Content network resources selection, path selection, QoS
Fig. 6. Generic Value Network Configuration (VNC) for content delivery in ICN
The major stakeholders that can take up the aforementioned roles in our scenario
are presented in Table 2. We also use parentheses to include the additional roles that
could be potentially taken up by stakeholders in other scenarios. Additionally, we
include the CDN providers, as well as the regulators that exist in current Internet,
although their interests and actions are not subject of this analysis.
4.3 Tussle Analysis

In this section we identify tussles related to key roles listed in Table 1. Each tussle is
described with references both to the use case (Fig. 5) and the VNC (Fig. 6).
Table 2. Stakeholder - basic role mapping
Stakeholder Basic role

End-user Content consumption, (content creation)
Content Provider (CP) Content creation, (content access management)
Internet Service Provider (ISP)
- Access Network Provider (ANP) Access network operation, cache management, cache
location ownership, content network management,
(name resolution , content access management)
- Inter-Connectivity Provider (ICP) Interconnectivity provisioning to ANPs,
(name resolution)
NRS provider Name resolution
Content Distribution Network Cache management, cache location ownership,
Provider (CDN), e.g. Akamai content network management, name resolution
Regulator Competition regulation
Tussles related to name resolution

Spam Requests Tussle: The local NRS may decide to replicate the requested
information to his own cache like the rendezvous in the pub/sub model. In this case,
the local NRS (or RENE) adds a subscription in his message towards the publisher
asking the information to be forwarded also to the ANP’s cache. Thus, an NRS could
issue a request for another stakeholder (e.g. the end-user) for an information item that
the latter is not interested in (spam). This combined service contradicts the
functionality separation as dictated in [2], since the rendezvous also performs content
management besides its main function, i.e., name resolution.
Net Neutrality Tussle: The global NRS is potentially in a position to favor specific
CPs by promoting their content over the content of other CPs, or by filtering the
information items provided by the latter ones. Additionally, if the local NRS is
provided by the ANP (similar to today’s ISPs’ DNS service bundled with access
provisioning), there is an incentive for the NRS to forward the subscription to the
local publisher. If the content is not locally published, then the ANP-owned local
NRS (NRS2) may refuse to further handle the request to avoid fetching the
information object from a remote publisher or the cache of a competing CDN to avoid
increasing ANP2’s interconnection costs. The latter case is also known as a “walled
garden”. Ideally this situation is avoided by having architectures that allow
competition in the resolution service; otherwise a regulator would have to ensure that
end-users are allowed to send their subscriptions to the NRS of their choice.
Conflicting Optimization Criteria Tussle: When multiple sources can serve a
request, a tussle occurs due to actors’ different preferences for the one to be used (e.g.,
cost concerns, performance attributes, regulatory constraints, or other local policies).
For example, localization of traffic due to caching and content replication affects the
volume exchanged between ANPs, as well as ANPs and ICPs. If the local NRS
forwards the content requests to local caches, both the interconnection costs of ANPs
and revenues of ICP decrease. This is naturally positive to ANPs but negative to ICPs.
Similarly, an ICP-owned global NRS may forward a subscription originated from a

local NRS to publishers that are located behind a transit link, even if the information
item was also available through a peering link (a different scenario than the one in
Fig 5). The same situation could appear if the local NRS is provided by a third-party,
similar to, e.g., Google’s DNS, which may have different incentives. Such conflicting
optimization criteria might imply a straightforward increase of interconnection cost
for the ANP, and possibly degraded end-users’ Quality of Experience (QoE).
As it is obvious, the actor who controls the name resolution is able to restrict or
even determine the available options to others. However, such an actor (like an ANP
when the end-user has used a different NRS provider) may still be able to use a
different source than the proposed one. For example in [6], after the final matching of
a publisher and a subscriber by the Rendezvous Network, the Topology Manager may
create a path between the subscriber and a different publisher (i.e., an ANP’s own
cache server)1. This could be the case when the end-user or the NRS provider cannot
verify which publisher has been actually used.
Furthermore, other stakeholders could enter the name resolution market. In an
extreme case, even a CP may react by providing also his own NRS. For example,
YouTube could serve its information space by redirecting end-users to servers according
to its own criteria). Such an NRS may also be provided as a premium service to other
CPs. However, in both cases, client configuration by the end-users is required.
Finally, traditional CDN providers (like Akamai) could also react by announcing
all the content items (publishers and caches) they are aware of to multiple NRS
providers, or even deploy their own name resolution servers.
Nevertheless, the name resolution role is central to ICN and of high interests to the
most stakeholders in this setup.
Tussles related to content access management
Access Control Tussle: If the ICN architecture does not clearly specify how to limit
access to certain end-users, the ANP may serve the subscriptions from its local cache
without consulting CP’s AAA system. This would destroy CP’s business, especially if
it is based on transactional payments from end-users, but also if he sells advertising or
information about content usage. A proposed solution is presented in [10], where the
RENE could act as an accountability broker between the end-users and CPs.
Content Usage Statistics Tussle: When the content is provided from local caches
controlled by multiple stakeholders, the CP may lose visibility on how its content is
used. This information has value, because payments from advertisers to CP and from
CP to content makers are often based on the popularity of content.
Privacy Tussle: Finally, a control tussle may rise between the stakeholder managing
content access and the end-users, since the former can use personal and transactional
data for purposes not approved by the end-user to make a profit, e.g. to sell data to
marketing companies.
1
Here, we assume that the Topology Manager is aware of the information item ID.
Tussle related to cache management

Content Freshness Tussle: The content cached in the ANP’s caches may be
outdated, because the ANP may be reluctant to update the content in order to reduce
his interconnection (i.e., transit) costs. Then, the end-user’s quality of experience
degrades, since he does not receive the most recent information.
Tussles related to cache location ownership
Cache Placement for Revisiting Interconnection Agreements Tussle: Tussles here
mostly involve ISPs since existing interconnection agreements may not be justifiable
if a new cache was added. Hence, ISPs may try to affect peering ratios in
advantageous ways (e.g. create such an imbalance that violates their peering
agreement). For example, an ANP deploying his own cache content network and
having a peering arrangement with another ANP (which does not own a content
network) may break this agreement in hopes of providing transit service to the latter
one. Similarly, an ICP who sees its revenues being reduced may decide to adjust
transit prices or enter the content delivery market by providing global NRS services.
Tussles related to content network management
Network Information Tussle: An ANP may provide inaccurate information (or no
information at all) about its network topology, dimensioning, current utilization, etc.,
fearing that this sensitive information could be revealed to its competitors. However,
this may have a negative impact on the effectiveness of selecting publishers and
consequently paths between publishers and end-users that meet the QoE constraints
posed by the latter. For example, in case there are two publishers for a particular
request, one of them may seem more appropriate (although it may not be), if its own
ISP is untruthful by providing biased network information (e.g. lower delay in a path).
5 Discussion
ICN brings new challenges in the Internet market, since name resolution services may
be offered by different stakeholders in order to meet their own optimizing criteria;
either by the ANP, or by a third-party (such as a search engine or a significant CP).
Such major stakeholders of today’s Internet are highly expected to extend their
activities to offer NRS’ in ICN.
Additionally, there is a crystal clear incentive for an ANP to deploy ICN, in order
to enter the content delivery market. Due to the information-oriented nature of the
network, an ANP could deploy his own caches, which implies that the ANP will gain
more control of the content delivery. Therefore, under suitable business agreements,
this will imply increase of his revenue, while simultaneously reducing his operational
costs due to more efficient content routing and reduction of the inter-domain traffic.
Moreover, CPs and end-users will also be affected; i.e. CPs will be able to provide
their content through more communication channels to their customers, while end-
users will enjoy increased Quality-of-Experience (QoE).
On the other hand, the emergence of ANP-owned CDNs will cause traditional
CDNs to lose revenues and control over the content delivery market. Thus, legacy
CDNs will probably react in order to maintain their large market share, or at least not
exit the market. CDNs may deploy their own backbone networks to interconnect their
own caches, but still they will probably not in position to deploy access networks to
reach the end-users; this is ANPs’ last frontier. Nevertheless, no matter how legacy
CDNs will react, such local CDNs owned by ANPs will (and already) be deployed
(e.g. At&T’s CDN). The evolution of this competition and the way that the system
will be lead to an equilibrium is the subject of future investigation and analysis.
Our contribution in this paper resides in the identification and analysis of tussles in
a generic ICN architecture, which should be considered by designers and engineers
that aim at deploying new content delivery schemes for the FI.
Acknowledgement. The authors would like to thank G. Xylomenos, G. D. Stamoulis,
G. Parisis, D. Kutscher C. Tsilopoulos and X. Vasilakos. The research of A.
Kostopoulos and I. Papafili has been co-financed by the European Union (European
Social Fund – ESF) and Greek national funds through the Operational Program
“Education and Lifelong Learning” of the National Strategic Reference Framework
(NSRF) - Research Funding Program: Heracleitus II-Investing in knowledge society
through the European Social Fund. C. Kalogiros is supported by the EU-FP7 SESERV
project. The research of T. Levä and N. Zhang is supported by the EU-FP7 SAIL
project. The research of D. Trossen is supported by the EU-FP7 PURSUIT project.
Open Access. This article is distributed under the terms of the Creative Commons Attribution
Noncommercial License which permits any noncommercial use, distribution, and reproduction
in any medium, provided the original author(s) and source are credited.
References
1. Clark, D.D., Lehr, W., Bauer, S., Faratin, P., Sami, R., Wroclawski, J.: Overlay Networks
and the Future of the Internet. Communications and Strategies 63, 109–129 (2006)
2. Clark, D.D., Wroclawski, J., Sollins, K.R., Braden, R.: Tussle in Cyberspace: Defining
Tomorrow’s Internet. IEEE/ ACM Trans. Networking 13(3), 462–475 (2005)
3. Kalogiros, C., Kostopoulos, A., Ford, A.: On Designing for Tussle: Future Internet in
Retrospect. In: Oliver, M., Sallent, S. (eds.) EUNICE 2009. LNCS, vol. 5733, pp. 98–107.
Springer, Heidelberg (2009)
4. Brown, I., Clark, D., Trossen, D.: Should Specific Values Be Embedde. The Internet
Architecture? In: ACM Context ReArch 2010 (December 2010)
5. EU FP7 SESERV project, http://www.seserv.org/
6. EU FP7 PURSUIT project, http://www.fp7-pursuit.eu/
7. EU FP7 SAIL project, http://www.sail-project.eu/
8. Kalogiros, C., Courcoubetis, C., Stamoulis, G.D., Boniface, M., Meyer, E.T., Waldburger,
M., Field, D., Stiller, B.: An Approach to Investigating Socio-economic Tussles Arising
from Building the Future Internet. In: Domingue, J., et al. (eds.) FIA 2011. LNCS,
vol. 6656, pp. 145–159. Springer, Heidelberg (2011)
9. Casey, T., Smura, T., Sorri, A.: Value Network Configurations in wireless local area
access. In: 9th Conference of Telecommunication, Media and Internet (2010)
10. Trossen, D., Kostopoulos, A.: Exploring the Tussle Space for Information-Centric
Networking. In: 39th TPRC, Arlington, VA (September 2011)
11. Sterman, J.: Business dynamics: Systems thinking and modeling for a complex world.
McGraw-Hill, Irwin (2000)
12. Named Data Networking (NDN) project, http://www.named-data.net/
13. Trossen, D., Sarela, M., Sollins, K.: Arguments for an Information-Centric
Internetworking Architecture. ACM Computer Communication Review (April 2010)
14. Pöyhönen, P., Stranberg, O. (eds.): (D-B.1) The Network of Information: Architecture and
Applications. SAIL project deliverable (2011)
A Systematic Approach for Closing
the Research to Standardization Gap
Bernard Sales, Emmanuel Darmois, Dimitri Papadimitriou, and Didier Bourse
Alcatel-Lucent Bell Labs

{bernard.sales,emmanuel.darmois,dimitri.papadimitriou,
didier_arnaud.bourse}@alcatel-lucent.com
Abstract. Standardization activities are recognized as one of the tools to

incubate research results and accelerate their transfer to innovative marketable
products and services. However, the European Commission (EC) research
community and its associated stakeholders acknowledge the lack of research
transfer via the standardization channel, generally referred to as the research-to-
standardization gap. This chapter analyzes the root causes for this gap and
proposes way forward. In particular research-focused standardization is
considered as the instrument to address this issue. This chapter shows that pre-
standardization should be supplemented by a methodology and its associated
process aiming to systematically analyze the standardization aspects of research
projects and by helping them out to draw their standardization strategy.
1 Introduction
The Digital Agenda for Europe [1] highlights the importance of ICT standards in
delivering interoperability between devices, applications, data repositories, services
and networks. It also stresses the fact that standards are to be used strategically as a
means of stimulating innovation and promoting interoperability of innovative
products.
In this context, the EC has published in June 2011 a series of measures with the
objective to have better standards for Europe and to have them faster [2]. As a follow-
up of the publication of the White Paper “Modernising ICT standardization in the EU
- The Way Forward” [3] and the related public consultation, one major requirement to
strengthen the system of standard-setting in Europe is the recognition that global ICT
standards will play a more prominent role in the EU, both from the standardization
strategy [4] and regulation standpoints. In particular, regarding EU funded research
projects, [4] states, e.g.: “Finally, standards can help to bridge the gap between
research and marketable products or services”. “A systematic approach to research,
innovation and standardisation should be adopted at European and national level to
improve the exploitation of research results, help best ideas to reach the market and
achieve wide market uptake.”
It is well recognized that standards are one important way to promote the
translation of research results into practical applications [3] [5] [6] and are also, in
A Systematic Approach for Closing the Research to Standardization Gap 19
certain circumstances, the necessary pre-condition for a large deployment and the
successful commercialization of a technology. However, research projects do often
not engage consistently in standardization because they are not yet convinced by the
benefits or/and return on investment of engagement or because they are not familiar
enough with their target standardization ecosystem or need guidance on how to
address the problem of what to do, where and when to promote their research results
in standardization. This lack of engagement is generally referred to as the research-to-
standardization gap. The need for practical pre-standardization framework to close
this gap is identified as a priority by all stakeholders, including research, ICT
industry, EC, but also the Future Internet Assembly (FIA - pre-standardization WG)
which has recently proposed a shared action plan to support standardization activities
[7] [8]. It is also well accepted that initiatives to better link ICT standardization and
ICT R&D appear to be most effective when carried out at the research planning phase
rather than simply at the execution phase of specific research projects [3].
Standardization awareness thus needs to be considered early in the research life cycle
and should be an integral part of strategic research agendas.
Starting in Section 2 with an informal survey on research projects requirements,
this chapter will analyze the following aspects of the standardization gap: i) what are
the root causes of the research-to-standardization gap, ii) how to cope with the
specifics of the standardization ecosystem compared to the usual scientific
environment and iii) how to satisfy the necessary conditions to efficiently transfer the
research results to standardization. For this purpose, Section 3 of this chapter
addresses the limits of the classical standardization process in case research results
need to be incubated in standardization. In this context, a research-focused
standardization phase (generally referred to as pre-standardization), feeding the
classical standardization process, needs to be put in place. However, pre-
standardization needs to be complemented by a methodology and its associated
process aiming to systematically analyze the standardization aspects of research
projects and by helping them out to draw their strategy. These aspects are discussed in
Section 4 of this chapter.
2 Learning from Research Projects
In order to start identifying potential expectations and requirements to address the

research-to-standardization gap from the research standpoint, an informal set of
discussions has been initiated with running Objectives 1.1 projects (UniverSelf,
ETICS, EARTH, MEDIEVAL, OneFIT, …) and specific EC Call 8 proposals under
preparation at the time of the writing of this chapter. A first list of requirements has
been collected and is presented in Table 1. It should be noted that the requirements
are written as received from the survey without any further interpretation.
Nevertheless if some of these requirements can not be implemented as such, they
translate needs from the research community.
20 B. Sales et al.
Table 1. List of Requirements

Requirements from research project Taxonomy/aspect of
the requirements
Provide a thorough analysis of standardization bodies candidates and propose Planning
specific steps and community-building activities.
Identify the gaps in current standardization to provide support in coordinating Planning
and informing about potential standardization topics.
Support/guidance during the project submission phase and project Planning/Guiding
contractualization phase to improve efficiency, coordination, return on
investment and to adapt the standardization plans -based on proposal's
topic(s), EC feedback (hearings/negotiation), reviews, execution phase,
exploitation phase...
Provide best current practices on which SDOs are suitable/appropriate targets Planning/Guiding
for given field of research/standardization, how to approach and provide
consideration in terms of agenda (slow start, SDO processes vs. project
timeframe).
Provide networking/connections/know-how/who's doing what on a list of key Guiding
standardization topics (and maybe have also special networking events, e.g.
FI weeks, ETSI workshops).
Document success stories which will be inspiring/leading examples (to Guiding
follow/repeat/adapt).
Make available a single, reference up-to-date knowledge base for current Guiding
status of standardization in given topics/areas vs. ending/running projects and
a brief (explanatory) history (why current situation, incl. wrt alternatives).
Better link the different instruments, e.g. EU-level: Clusters, FIA, projects Linking
and SDOs, as well as at worldwide level incl. North America, Japan/Asia
initiatives/instruments.
Support/help after the end of projects to continue/follow-up initiated Following-up
standardization actions.
Provide up-to-date information with respect to the discussions in the Mutualizing
standardization bodies and support to the submission of contributions from
projects in a coordinated way
Create a platform for joint contributions and submissions to international Mutualizing
standardization bodies.
Set-up of an open forum where participants from different SDO could meet, Mutualizing
avoiding the need for some partner to attend ETSI, IETF, 3GPP meetings to
grasp the key challenges faced by operators and manufacturers.
These initial requirements and their taxonomy are a good starting point to frame the
discussion on what is needed to address the research-to-standardization gap, for
instance:
• Regarding the second requirement in Table 1, the identification of the gaps
should be in close communication with other standardization stakeholders (the
industry, regulators, standardization bodies) since researchers on their own are in
a bad position to identify the gaps effectively.
• The requirement to "make available a single reference up-to-date knowledge base"
seems difficult to achieve but however, [9] provides a first step in this direction.
• The requirement on "support/help after the end of projects to continue/follow-up
initiated standardization actions" is really crucial since without such support,
standardization plans in the typical short-lived research projects might not be
achieved, especially in cases the standardization eco-system is not ready to
progress the standardization objectives of the project.
It is anticipated that the “Planning” and “Guiding” aspects are necessary conditions to
reduce the research-to-standardization gap (note that the COPRAS project conducted
in the context of FP6 [10] took the same assumptions). On the other hand, the
“Linking”, “Following-up” and “Mutualizing” aspects provide means to support more
efficiently the pre-standardization actions. As one of the objectives is to address the
root causes of the research-to-standardization gap, the focus of this chapter is placed
on “Planning” and on “Guiding” while the “Linking”, “Following-up” and
“Mutualizing” aspects will no longer be discussed in this chapter.
3 Standardization and Pre-standardization
Standardization of protocols and interfaces has played and is still playing a key role in
the Internet development. In particular, the IETF has imposed itself as the main
Internet protocols factory while other standardization bodies like IEEE, ITU-T, 3GPP
and W3C are standardizing the infrastructure and technology enablers creating the
necessary open ecosystem that contributed to the Internet development.
However, the work in standardization is dwindled by its participant strategy in
terms of R&D and conflicting business objectives leaving in practice a very little
window to the research and academia communities to influence the process. One
could observe that in the early days of the Internet, its standardization was driven by
the research community. This materialized by the creation of the IETF that was an
emanation from the research community. Over time, as the Internet and its associated
technologies progressively matured and were deployed at a larger scale, the Internet
standardization gradually shifted to engineering and operational problems (the IETF
is often qualified today as "problem-driven"). As a result, even though the research
community is still involved in the Internet standardization process, its influence is
eroding over time. Nevertheless, the involvement of the research community in
standardization can bring a lot of added value to the industry (in particular when
practical use cases are identified at this stage of the process) since it allows early de-
risking of disruptive ideas by confronting them to i) executability/developability, ii)
deployability, and iii) market environment and, if successful, will accelerate
penetration of those innovative ideas.
In this context, a research-focused standardization phase needs to complement the
classical standardization process. In this model, the research-focused standardization
phase will feed the classical standardization process with a stream of de-risked ideas
that will, if successful, lead to a full standardized solution. It has to be noted that the
interactions and discussions in the context of pre-standardization can also directly
feed back the research project with valuable inputs to be further considered inside the
project (“external loop”). For this reason, this phase intends to bridge the research-to-
standardization gap and is generally referred to as the pre-standardization phase.
Major standardization bodies are adapting their processes to capture these
requirements. For instance, ISOC created in the 90’s the IRTF (the research arm of
the IETF), the ITU-T defined the concept of Focus Group, the IEEE established
IEEE-SA Industry Connections Program and the W3C the W3C Incubator Activity.
22 B. Sales et al.
In 2006, ETSI defined the concept of Industry Specification Group (ISG). All these
pre-standardization processes share the same principles: they are open to academia
and are based on a lightweight procedural structure compared to their “mother”
standardization groups. On the other hand, one can observe that, in the context of the
Internet these pre-standardization structures are not yet used at their full potential. In
particular, when "pre-standardization" processes/organizations exist, they have often
evolved in two directions, either by focusing on shorter-term engineering problems
the standardization body is recognized for (and, in turn, being perceived as no longer
fulfilling a research role) or by focusing on longer-term architectural problems (and,
in turn, being perceived as disconnected from the rest of the standardization
organization activities). It is also anticipated that the results of the Future Internet and
Future Networks research will have the potential to boost the volume of pre-
standardization activities and could really lead to the launching of the Future Internet
pre-standardization process.
It should be noted that not all research results need to be incubated in pre-
standardization. Depending on the standardization lifecycle and rationality, research
results can go directly to the classical standardization regime without going through a
preliminary pre-standardization phase.
For instance, the classical standardization regime is not yet ready to standardize all
aspects related to Self Managed Networks and, as a result, pre-standardization is
required (e.g. in an ETSI ISG or in an IRTF Research Group). In contrast, regarding
Carrier Ethernet, the standardization regime is mature; there is no need to go through
a pre-standardization phase.
Pre-standardization is the necessary tool helping create an environment that is,
when required, more suitable to incubate research ideas than the classical
standardization regime. Despite its great potential, pre-standardization alone (i.e.
without a built-in link to standardization and without a framework to systematically
analyze the standardization aspects of research projects and helping them out to draw
their strategy) is not broad enough to motivate researchers to present and defend their
ideas only there.
4 Methodological Aspects
4.1 The Need for Standardization Strategy
According to the experience acquired over years by the co-authors of this chapter, in
order to be really effective, standardization actions should be defined from and
supported by a well defined standardization strategy/planning. In the context of this
chapter, a standardization strategy is defined as a path of standardization-related
actions and objectives (in a few complex cases, a strategy may even comprise parallel
paths). Without any standardization strategy, the standardization actions are in general
unsuccessful or lead to suboptimal results. In the worst case, the standardization
achievements may even be conflicting with the research objectives of the project.
As a result, the standardization strategy has to be carefully addressed and has to

consider multiple dimensions including the maturity of the standardization ecosystem,
the position of the technology proposed for standard in the standardization life-cycle,
the objectives of the research projects, the possible open issues and the research
project maturity. To deal with the dynamics of both standardization and research
environments, the strategy needs to be reassessed on a regular basis. As a
consequence, this process is characterized by iterative cycles of defining/refining the
strategy, adapting actions and expected achievements in standardization bodies.
To help and guide research projects, a methodology to analyze the standardization
aspects of research projects and the associated process (mechanisms) need to
be further developed. The combination of the proposed methodology and its
associated process will enable the research projects to define and reassess their
standardization strategy that is a necessary condition to address the research-
to-standardization gap.
4.2 A Systematic Way to Draw the Strategy

A systematic methodology to analyze the standardization aspects of a research project
and to draw its standardization strategy/approach was already proposed by the authors
of this chapter [11] [12]:
• Step 1: Identify what needs to be standardized (interfaces, etc) to allow the

technology proposed by the project to be interoperable and deployable at large
scale. In general, this step implies the identification of an “initial” architecture.
• Step 2: Identify the role and impacts of standardization bodies on the business
segment targeted by the project. At this step, standardization bodies are categorized
as fulfilling a role in the standardization food chain, i.e. requirements, architecture,
solution/protocol/interface and interoperability/testing.
• Step 3: Evaluate the need to improve the standardization eco-system to maximize
the chance of success, this can materialize either by creating new (pre-)
standardization technical committee and/or by attracting major stakeholders.
• Step 4: Identify the “structuring” dimensions (i.e. what characterizes the
standardization objectives trajectory/path) for the proposed technology/system to
define a) the criteria to shape the associated standardization target(s) of the
research projects b) the necessary conditions to meet in order for the
technology/system to enable its standardization. The output of this step is a
standardization objectives trajectory to be realized.
The main objective of the methodology is to guide research projects in identifying

their standardization needs and approach in a systematic way to ensure that all the
necessary aspects are analyzed and developed. The methodology can lead to three
types of results:
24 B. Sales et al.
1. The first one is when standardization is not needed at all and when the lack of
standardization is not a roadblock for the large scale deployment of the
technology being designed by the research project.
2. The second case is when standardization is required but the related standardization
ecosystem is not ready/in place to progress the standardization objectives of the
project. In other words, it is very unlikely that a standardization body will accept
to incorporate this necessary work items in its standardization work program. In
this case, the technology needs to be incubated in a (pre-)standardization group. In
general, this will require the creation of a new (pre-)standardization WG.
3. The latter case is when the technology can be directly pushed in standardization
bodies without the need to go through a pre-standardization phase.
In general, Step 1 of the methodology is conducted in an analytical way. In this

context, having an (initial) architecture is of a great benefit since it will enable to
systematically enumerate all the interfaces and to analyze formally which of them
needs to be standardized to enable the further transfer of the technology/system to
marketable products and/or services. This analytical study can be complemented by
an experimental facility/test-bed whose objectives, when affordable, are to benchmark
components, to identify their behavior in large-scale setup, and to detect non
interoperable components/features that will at the end require some form of
standardization. It should be noted that this four steps methodology has been already
successfully applied to several Alcatel-Lucent small and medium size research
projects. This methodology was also used to define the standardization plan of some
EC FP7 Future Internet research projects. This contributes to validate the applicability
of the methodology but does not demonstrate that the methodology can be deployed at
the large scale (e.g. at the FIA level).
4.3 Application of the Four Steps Methodology to ECODE Project

FP7-ICT 2007 ECODE project (Grant 223936) [13] is the first FP7 project to which
the above methodology was used.
The objective of the ECODE project was to associate new architectural network
components, based on machine learning principles, architectures and techniques in
networking platforms to assist operation (automated, on-line analysis), improve
performance gain by predicting and adapting decisions, and extend Internet
functionality (e.g., diagnostics, network intrusion/attack detection, etc.) [14].
The four-step methodology was applied at the time of writing the project proposal;
the application of the methodology was employed to document the standardization
part of the project proposal. The results of the initial application of the four steps of
the methodology was summarized in the form of a dashboard [12] that was used at
different stages of the project (see Figure 1).
Fig. 1. Standardization Dashboard for Research Projects
In this context, Step 4 of the methodology enabled the ECODE team to link their
envisioned usage scenarios for their technology (identified as structuring aspects by
the methodology) with specific standardization objectives. In particular by applying
the step 4 of the methodology, the ECODE project has identified two objectives
for the introduction of machine learning component (the above-defined ‘structuring’
aspects): 1) address current Internet operational challenges; 2) further extend Internet
functionalities (diagnosability, security, etc.).
From the standardization perspective, the first objective implies that protocols must
be standardized in the IETF, while the second one implies that an advanced
architecture should be defined, e.g. in ETSI. In addition, as machine learning
techniques were never used before in the context of Internet and are challenging in the
context of the Internet deployment, it would be necessary to have a pre-
standardization phase, e.g. in the IRTF (see the dashboard in Figure 1).
Using the proposed methodology, the standardization strategy was reassessed twice
in the course of the project due among others to the change in IRTF priorities. This
reassessment helped the ECODE partners in determining their standardization plan
beyond the lifetime of the project. All these steps enabled the ECODE project to
define and refine systematically a coherent standardization strategy starting from
requirements, followed by the identification of the target standardization bodies and
roles and ending with the definition of the standardization approach and objectives.
4.4 Implementing the Methodology at Large Scale

Having outlined a methodology to assist the research community to identify the
standards needs, approach and objectives associated to a research project, the
“process” aspects (the mechanics) that will enable the implementation and validation
of the methodology on a large scale (i.e. at the FIA level) have to be addressed.
26 B. Sales et al.
For this purpose, the relation between research projects and their standardization
ecosystem are analyzed in terms of downstream and upstream channels. The
downstream channel is materialized by the participation and contributive efforts of
research experts to the standardization bodies: participation to meetings, submission and
presentation of contributions and leadership positions when appropriate. This
downstream channel is generally managed by the research project, resulting in a
standardization approach defined at the project level as part of the dissemination and
exploitation plans. However, as already mentioned before, researchers are not
necessarily attracted by or familiar with the targeted standardization environment. There
are multiple reasons that can explain this:
1. Research project objectives are research results-driven whereas
standardization objectives are engineering consensus-driven.
2. Participation to most standardization bodies requires an annual fee. Unless
that cost can be sustained by academic and research institutes, the research
project cannot access the standardization organizations working documents
(contributions, meeting minutes, etc.);
3. Standardization debates and positioning of actors are often driven by
"economical" interests beyond any possible influence of academics and
research institutes (not recognized as full-fledged players).
4. Each standardization body operates with its own specific methods and
procedures whilst research projects standardization plans require combining
actions in multiple standardization bodies - which in turn increase the
complexity for the research project to conduct its standardization actions.
As a result, the standardization strategies and plans of a research project are often
defined on an ad-hoc basis and sometimes, even misleading and/or incomplete. When
a project has an insufficient understanding of the standardization environment, it may
opt for easily implementable workarounds. For instance, its contributions are
submitted only once to a standardization organization and sometimes not presented in
meetings. In this case, the standardization body just “notes” that the contribution was
submitted and, as a result, the technology designed by the research project will never
lead to a standard. Moreover, contributions from research projects are also often
missing their target: expecting that the outcomes of research as reported in project
deliverables will be accounted for as-is by the targeted standardization organization is
not realistic. Two main causes for failure can be identified: i) lack of adoption of the
conventions and writing style of the targeted standardization body, and ii) difficulty
to confront its output with various technical communities (system engineers, network
engineers, operation, etc.) before it can have a technological impact on the course of
the standards making.
In addition to the ‘downstream’ channel, there is also an ‘upstream channel’ from
the standardization community to the research projects. In the simplest way currently
available, this corresponds to the information published by standardization
organizations on their web sites. This information is often general purpose and as
such not targeted and/or tailored to/for the research community; it is at best
informative but often rather useless for researchers. As noted, if project partners do
not pay the standards organization membership fees (when applicable), this
information is not even accessible at all (e.g., for copyright reasons). In some cases,
this upstream channel is better managed when standardization bodies organize
‘research to standardization’ workshops (e.g. [5], [6]) though, often, the audience on
these workshops is composed of the research experts already involved in the
standardization work.
It is postulated by the authors of this chapter that three conditions need to be
satisfied in order to improve the quality of the downstream (from research to
standardization) channel and maximize the value of the output: 1) availability of
information from standardization bodies that is directly relevant to the research
project; 2) mutual understanding, at both ends of the channel, that research results
have reasonable chances to be adopted in the appropriate standardization context; and
3) joint determination of the trajectory (sequence of standardization actions with
starting and ending points) by means of a standardization strategy.
To satisfy the three conditions to improve the downstream (from research to
standardization) channel, the upstream (from standardization to research) channel
needs to be enhanced in the following ways:
1. Provide information related to standardization status and evolution specifically
targeted to the research community. (A first step in this direction is the
information repository provided by the FIA pre-standardization WG [9])
A criterion of success for this approach is the initiation, within a
standardization organization that takes this path, of a standardization track that
was not previously addressed. Two cases shall be however distinguished. In the
case of a standardization organization already working on the technology to
which the research project contributes, it is less complex to put in place the
process, but the impact on the technology specification will probably be
smaller. When the standardization organization is not yet working on the new
technology proposed by the research project, more effort will be required but -
in case of success - impact will be greater since it will define a new technology
specification track.
2. Proactively support the research project by a team of dedicated experts with a
strong ‘research and standardization’ background. The role of these experts, the
‘Research-to-Standards’ team, is i) to guide the research projects on the
definition of their standardization strategy (using the methodology defined in
Section 6) including the sequence of standardization actions required to ensure
that the technology under consideration will be developable and deployable at
a large scale (necessary condition), and ii) to regularly follow-up with research
teams on progress and open issues and/or blocking factors, to help progressing
on the trajectory and propose possible remediation actions in case of problem.
3. Research projects must be convinced of the benefits to use a well defined
methodology to define their standardization strategy and trained on how to use
the methodology.
In the context of autonomic networking (e.g. see [15]), the downstream channel from
research projects to standardization is currently working quite well e.g. in terms of i)
28 B. Sales et al.
number of contributions, ii) stepwise approach based on architecture – uses cases –

solutions, iii) improvement of the (pre-)standardization infrastructure with the
creation of the ETSI ISG on Autonomic network engineering for the self-managing
Future Internet (AFI), iv) reach in terms of standardization bodies, etc. The only issue
is the critical mass, i.e. only few FP7 research projects dealing with autonomic
networking are involved in the (pre-)standardization process. If someone had
implemented the improvements proposed in this Chapter to the upstream channel
from standardization to the research projects dealing with autonomic networking, the
expected results would have been to embark in the standardization effort almost all
the running FP7 research projects relevant in this context (and even EUREKA and
National projects in case of full and well organized implementation of the proposed
process).
Regarding the ECODE project discussed in section 5.2, the downstream channel
was not working as expected (i.e. the creation of an IRTF Research Group), but this is
mainly due to the change of priorities in the targeted pre-standardization body.
However, the enhancements to the upstream channel as proposed in this chapter
enable the project to adapt their standardization strategy to cope with this situation.
Future work will consist in applying the methodology on a set of representative
research projects in order to characterize the expected benefits and give more
guidelines and cook book on its implementation.
5 Conclusion
Research-focus standardization (in general referred to as “pre-standardization”) is a
necessary instrument to attract a critical mass of researchers to participate in
standardization process. But this instrument alone is not sufficient. Actually pre-
standardization should be supplemented by a dedicated planning effort at the project
research level that will have to be materialized in a well defined standardization
strategy. However, standardization body operates with its own specific methods and
procedures. In addition, the necessary research projects standardization actions
require combining actions in multiple standardization bodies which in turn increase
the complexity for the research project to define its standardization strategy. As a way
to guide the research projects, the authors provide a methodology and its associated
process aiming to systematically analyze the standardization aspects of a project and
by helping them out to draw their strategy.
The above enhancements can be either implemented by key representative
standardization organizations or implemented by an entity external to standardization
bodies (but closely linked/interacting with the key standardization organizations).
To adopt these enhancements, standardization bodies must be convinced of the
usefulness of the approach before engaging resources to implement the proposed
process. It is currently difficult to anticipate the benefits of having this process
implemented in key standardization organizations or in an entity outside the
standardization bodies. Even more important, research projects must be convinced of
the benefits to use a well defined methodology to define their standardization strategy
and should be trained on how to use the methodology. The authors believe that the
proposed process, once validated in the Future Internet context, e.g. on a selected
representative research projects, can be deployed at a large scale and deliver the
expected benefits to research and standardization.
References
1. European Commission, A Digital Agenda for Europe, COM(2010) 245 final/2, Brussels
(August 26, 2010)
2. European standardisation policy, http://ec.europa.eu/enterprise/
policies/european-standards/standardisation-policy/index_en.
htm
3. European Commission, Modernising ICT Standardisation in the EU: the Way Forward,
COM(2009) 324, Brussels, July 3 (2009)
4. European Commission, A strategic vision for European standards: Moving forward to
enhance and accelerate the sustainable growth of the European economy by 2020, COM
2011, 31, Brussels, June 1 (2011)
5. ETSI, Future Network Technologies Workshop, Sophia Antipolis (March 10-11, 2010),
http://www.etsi.org/WebSite/NewsandEvents/Past_Events/201_Fu
tureNetworkTechnologies_WS.aspx
6. ETSI, 2nd ETSI Workshop on Future Networks Technologies, Sophia Antipolis
(September 26-27, 2011), http://www.etsi.org/WebSite/NewsandEvents/
Past_Events/2010_FutureNetworkTechnologies_WS.aspx
7. Bourse, D.: Future Internet Assembly (FIA) Pre-standardisation WG - Review of
Objectives and Progresses, Poznan FIA, Poznan (October 25, 2011)
8. FIA Standardisation Support, http://fisa.future-internet.eu/index.php/
FIA_Standardisation_Support
9. Pre-standardisation activities in FIA, http://fisa.future-internet.eu/
index.php/Pre-standardisation_activities_in_FIA
10. COPRAS project, http://www.w3.org/2004/copras/
11. Papadimitriou, D., Sales, B.: A path towards strong architectural foundation for the
internet design. In: 2nd Future Network Technologies Workshop, ETSI, Sophia Antipolis,
France, September 26-27 (2011)
12. Papadimitriou, D., Sales, B.: Cognitive Augmented Routing System and its
Standardisation Path. In: Future Network Technologies Workshop, ETSI, Sophia
Antipolis, France, March 10-11 (2010)
13. ECODE (Experimental COgnitive Distributed Engine) project, http://www.ecode-
project.eu/
14. Papadimitriou, D., Donnet, B.: A Cognitive Routing System for the Internet. In: 8th
Würzburg Workshop on IP (Euroview 2008): Joint EuroNF, ITC, and ITG Workshop on
Visions of Future Generation Networks, Würzburg, Germany, July 21-22 (2008)
15. Ciavaglia, L., Altman, Z., Patouni, E., Kaloxylos, A., Alonistioti, N., Tsagkaris, K.,
Vlacheas, P., Demestichas, P.: Coordination of Self-Organizing Network Mechanisms:
Framework and Enablers. In: Proceedings of the Special Session on Future Research
Directions at ICST MONAMI 2011 Conference, Aveiro, Portugal (September 2011)
SOCIETIES: Where Pervasive Meets Social
Kevin Doolin1, Ioanna Roussaki2, Mark Roddy1, Nikos Kalatzis2,

Elizabeth Papadopoulou3, Nick Taylor3, Nicolas Liampotis2, David McKitterick4,
Edel Jennings1, and Pavlos Kosmides2
1
TSSG, Waterford Institute of Technology, Waterford, Ireland
2
National Technical University of Athens, Athens, Greece
3
Heriot-Watt University, Edinburgh, UK
4
INTEL Labs Europe, Leixlip, County Kildare, Ireland
Abstract. Traditionally, pervasive systems are designed with a focus on the

individual, offering services that take advantage of their physical environment
and provide a context-aware, personalised user experience. On the other hand,
social computing is centred around the notion of a community, leveraging the
information about the users and their social relationships, connecting them
together often using different criteria that can range from a user’s physical
location and activity to personal interests and past experiences. The
SOCIETIES Integrated Project attempts to bridge these different technologies
in a unified platform allowing individuals to utilise pervasive services in a
community sphere. SOCIETIES aims to use community driven context
awareness, preference learning and privacy protection for intelligently
connecting people, communities and things. Thus, the goal of SOCIETIES is to
radically improve the utility of Future Internet services by combining the
benefits of pervasive systems with these of social computing. This paper
provides an overview of the vision, concepts, methodology, architecture and
initial evaluation results towards the accomplishment of this goal.
Keywords: Pervasive Computing, Social Networking, Pervasive Communities,

Cooperating Smart Spaces, Community Interaction Spaces, Future Internet.
1 Introduction
Pervasive computing [1] is the next generation paradigm in computer science that
aims to assist users in their everyday tasks in a seamless unobtrusive manner, by
transparently and ubiquitously embedding numerous computing, communication and
sensing resources in the users’ environment and devices. Until now, pervasive
computing systems have been designed mainly to address the needs of individual
users. This neglects an important part of human behaviour; socialising, and might
partly explain the slow take-up of pervasiveness in commercial products. On the other
hand, social computing [2] has enjoyed meteoric success in bringing people together
online. Products in this area, however, do not integrate well with any but a few of the
many devices and services to which their users have access.
This paper describes the work being carried out in the FP7 SOCIETIES (Self
Orchestrating Community Ambient Intelligence Spaces) integrated project (www.ict-
societies.eu), the aim of which is to investigate and address the gap between pervasive
SOCIETIES: Where Pervasive Meets Social 31
and social computing by designing, implementing and evaluating an open scalable

service architecture and platform for our so-called Pervasive Communities. Pervasive
Communities have the potential to completely transform traditional online social
networks, freeing them from web-applications and letting them loose in the real
physical world. SOCIETIES supports the creation of purpose-driven Pervasive
Communities by finding, connecting and organising relevant people and things from
both physical and digital environments.
The core value proposition of Pervasive Communities is in continuous evaluation
and refinement based on feedback collected from three real user trial groups that are
involved in the project from the start and will be engaged until the project’s
completion. These user groups are: (i) the Student community, consisting of students
from Heriot-Watt University in Edinburgh, (ii) the Disaster Management community
of experts from the European Civil Protection Mechanism and (iii) the Enterprise
community from Intel's offices in Ireland.
The purpose of this paper is to elaborate on the vision, concepts, methodology,
architecture and initial evaluation results towards the realisation of Pervasive
Communities. The rest of this paper is structured as follows: in Section 2, the
SOCIETIES vision is presented and the concepts introduced are defined. Section 3
elaborates on the research challenges being investigated towards the Future Internet. In
Section 4, the methodology adopted is described spanning from user research
techniques, to technical requirement extraction methods and business analysis tools
employed. Section 5 presents the architecture that has been designed and is being
implemented, which exploits the benefits of both the pervasive computing and the social
computing paradigms. In Section 6, the initial evaluation methodology and results are
described. Finally, in Section 7, conclusions are drawn and future plans are exposed.
2 Vision and Concepts
2.1 The SOCIETIES Vision

While the majority of human social interaction takes place in the physical world, the
digital world is becoming increasingly integrated into the social fabric in which we
(co-)exist. The overarching goal of the EU funded project SOCIETIES is to
seamlessly integrate the social aspects of our physical world with our digital
equivalents.
The focus here is on the Discovery, Connection and Organisation of people and
things (sensors, context data, devices, resources, services, information, etc.) into
organised dynamically formed pervasive communities which provide their members
with an enriched social experience, supported by enhanced proactive behaviour. The
terms “discover”, “connect” and “organise” describe the full functionality that
supports the entire lifecycle of pervasive communities.
Discovery refers to the finding of people, communities, services, devices and
resources across the physical and virtual worlds. What is critical to note here is that
SOCIETIES provides the capability to discover entities that are relevant to a
particular situation (be it a user’s goals, desires, current context, etc.). This relevant
discovery enables the specification of deep associations between a group of entities,
with an external interface being made available in order to provide a third party
32 K. Doolin et al.
service provider, or simply a user, with access to that interconnected community. This
pushes beyond the capabilities of current social networks and services which rely
heavily on, for example, static personal information and user preferences, or manually
provided context changes (such as a manual check-in). This allows for the provision
of intelligent, rich, contextual data about users and the entities they interact with.
Fig. 1. The SOCIETIES vision
Harnessing these discovered entities, SOCIETIES then enables their connection

across the physical and virtual worlds, allowing them to communicate and to capital-
ise on the capabilities of the entities each has formed a connection with as part of the
pervasive community. This allows services such as crowdsensing, to take advantage
of the multiple devices of the many members of a community and can provide them
all with a much richer contextual picture of their physical and digital worlds than any
one of them could generate alone. Trust, privacy and security are critical factors being
considered in all phases of the project.
With connections in place, SOCIETIES can progress to organise and manage the
lifecycle of this dynamically established community. This lifecycle management
includes the introduction/removal of entities to/from the community – based on
further discovery and connection cycles – and the formation of super-/sub-
communities based on the context of the community as a singular entity. The results
of the organisation allow for the creation of new service paradigms that use a
community of connected entities as a focal point. The realization of pervasive
community Discovery, Connection and Organisation in this fashion requires a
significant effort in terms of research and development of new innovative
functionalities, many of which provide value for third parties either as standalone or
as more integrated intelligent features.
2.2 Concept Definitions

To address the vision above, the notion of Cooperating Smart Spaces (CSSs) has been
introduced that aims to extend pervasive systems beyond the individual to dynamic
communities of users. CSSs enable groups of users that demonstrate commonalities
for a non-trivial period to join together in pervasive communities. To enable the
merging of pervasive and social computing, in addition to the CSS notion, several
new concepts have been introduced. These concepts are defined below.
A Pervasive Community is a group of, two or more, individuals who have agreed
to share some, but not necessarily all, of their pervasive resources with other members
of that community. The Pervasive Resources that can be shared are: (i) services,
including services for controlling personal and environmental devices and (ii)
information (both individual and community), including context, preferences,
behaviours and memberships. A pervasive community, once constituted, forms a
Community Interaction Space (CIS). There is a one-to-one mapping between
pervasive communities and CISs. Individuals may belong to any number of pervasive
communities, and thus CISs, simultaneously.
Members of a pervasive community interact with a CIS via their own personal
Cooperating Smart Space (CSS). CSSs create the building blocks for enabling the
integration of pervasive computing with social communities (physical or digital).
CSSs constitute the bridge between a user's context (devices, sensors etc.) and the
community the user is a part of. A CSS is a digital representation of a user or
organisation, and also defines the impact that their services, information and resources
have within a set of communities. As such, it represents the user’s dynamic
contribution to one or more communities. The CSS provides its owner with a suite of
services which support the creation of, and participation in, pervasive communities as
well as a range of intelligent cross-community functionalities, which enable the
individual community member to benefit from the information and services of the
community as a whole. A community is a collection of CSSs and/or supporting
infrastructure services that wish to collaborate for mutually agreed purposes. There is
a one-to-one mapping between individuals and CSSs. The only way in which an
individual can participate in a CIS is via their CSS, but they can also interact with
other CSSs without having to form pervasive communities or create CISs. Individuals
may also interact with other individuals without using CSSs at all by employing more
traditional mechanisms.
3 Research Challenges towards the Future Internet

SOCIETIES provides a platform that enables individuals to connect their physical
activities with their online social environment in a transparent manner. Users avail
themselves of the pervasive features of CSSs and CISs, such as context-aware,
proactive and personalised service adaptation, implicit & explicit user behaviour
learning, personalised privacy protection and cross domain access to services &
resources. On an individual basis, pervasive information, such as user behaviour
models and context information, are enhanced with knowledge inferred by monitoring
the social interactions of the user on the social networks they visit. Augmenting the
personalisation and context information improves the quality of the pervasive
technologies that CSSs offer. On a community level, SOCIETIES utilises data such as
user behaviour, context and trust to form communities of users who share similar
interests, hobbies, careers, etc., enabling free exchange of information and
knowledge. Moreover, it provides an open scalable service architecture that allows
users to share resources (e.g., services, devices, sensors) with members of the same
community equipped with a range of resource sharing policies for resolving conflicts.
Furthermore, services and resources can be configured at both CSS and CIS level by
taking into account community context and behaviour that is inferred by collecting
and mining information from the members of the community.
34 K. Doolin et al.
The most essential resource for realising the CSS vision is the availability of
information and more importantly personal information. To address the issue of
privacy that rises, CSSs provide a range of intelligent privacy protection techniques
for managing the flow of information and allowing the user to have complete control
over the handling and disclosure of their information. Users can explicitly create
privacy preferences that state how their information is disclosed while the system is
also able to implicitly learn privacy preferences by monitoring the user’s behaviour
related to privacy protection.
Clearly SOCIETIES draws together a number of key challenges for the Future
Internet. Social computing, in many different contexts and through various devices, is
becoming a major driver for Internet use. Pervasive systems, as embodied in smart
spaces, are also set for a deployment explosion and will capitalise on the Internet of
Things to make enormous demands on the Internet in the near future. The social and the
pervasive aspects of Internet use each raise important privacy challenges in their own
right but together, the risks and consequences of failing to provide adequate and usable
privacy mechanisms increase exponentially. In addressing all of these challenges
SOCIETIES is making a significant step towards shaping the Future Internet.
4 Methodology
4.1 User Research Methodology

The user research methodology adopted was required to meet the following
objectives: (i) introduce users to CSS concepts and novel technologies, (ii) motivate
users to participate through engaging with manifestations of CSS visions that are
meaningful to them, (iii) yield requirements from user research that would be initially
informative to project development, (iv) enable researchers to have some useful and
empathetic insight into the lives, interests, and concerns of potential users, (v) utilise
research activities that could be conducted in a minimal amount of time with little
effort or disruption to the lives of the users and (vi) facilitate equal access interaction
between the stakeholders of user groups and researchers with malleable scenarios.
The user research methodology employed a triangulation of methods (Figure 2) that
included: observation via ethnographic methods, self-reporting via online surveys and
scenario led participatory workshops. For each group, the approaches used varied taking
into consideration the location, environment, user access and availability. Results from
the three different approaches of user research were presented as ethnographic vignettes,
statistical analysis and updated scenarios. User requirements were extracted from
combining these three sets of results for each of the three user groups.
Rapid ethnographic techniques, such as participant observation, contextual
enquiry, shadowing and guerilla fieldwork were employed to observe people in their
natural environment, thus gaining insights and understanding about the everyday
worlds of users in each group. The results of this research were presented as ethno-
graphic vignettes and field observations.
An online questionnaire was formulated to gather information about each user
group’s general demographics; their knowledge and experience of technical tools,
pervasive features, and social networking services; their current community
connections, and other relevant information specific to each group. The survey was
Fig. 2. User Research Methodology
sent to a random sample from each group. The responses were anonymised. Graphs of
the results demonstrate commonalities and differences in the social use of technology
within and across the groups.
Participatory Design (PD) workshops were organized to provide a democratic,
collaborative approach, facilitating creative, cooperative involvement of all the stake-
holders in the development of project concepts and services. Scenarios were selected
as a key tool for the PD sessions, as they function both, as a creative process for
visioning exercises, as well as an empathetic narrative conduit for complex ideas and
information. Initial scenarios demonstrating possible uses of the proposed platform in
the context of student, enterprise and disaster management situations were sketched in
brainstorming sessions with researchers. These initial scenarios were in turn
introduced to users by researchers in the neutral creative third space of PD workshops
[3], where participants’ reactions, ideas and discussions led to alterations and
advancements of these scenarios. Creative understandings [4] forged in these sessions
led to updated scenarios envisioning how pervasive communities could function in
each group’s social setting.
4.2 Technical Requirement Extraction Methodology

Based on an evaluation of the state of the art methodology approaches, it has been
decided to classify the technical requirements in a manner similar to the approach
suggested by the FURPS model [5]. More specifically, a scenario-driven process has
been followed to collect and specify the technical requirements. The five stages of
this elicitation process are illustrated in Figure 3 and are briefly described below.
36 K. Doolin et al.
Fig. 3. Technical Requirements specification methodology
1. Scenario brainstorming: This stage aimed to the design of various story flows and
scenes demonstrating and extending the features of the envisaged system.
2. Gathering of initial requirements: In this stage, an initial set of functional
requirements was extracted from the scenarios produced in stage 1. These
requirements were classified in five main categories: General, Deployment,
Service and Resource, User Experience and Security-related Requirements.
3. Scenario evaluation, analysis, ranking, filtering and refinement: In this stage, the
initial scenarios produced in stage 1 were evaluated and ranked based on various
criteria, such as the volume of features they demonstrated compared to the feature
set captured in the vision of Section 2, the quality of the initial requirements
collected in the second stage, etc. Based on this evaluation, on the end-user
feedback collected (§ 4.1) and on the business analysis performed (§ 4.3), a set of
refined final scenarios was produced.
4. Refinement of functional and non-functional requirements and extraction of use
cases: In this stage, the final scenarios have been studied in order to extract
additional technical requirements (both functional & non-functional), as well as
use-cases, while the initial requirements collected in stage 2 were homogenised,
merged, eliminated, extended and classified.
5. Harmonisation, prioritisation and ranking of requirements: In the final stage, the
elicited requirements were prioritized, harmonised and checked for consistency.
4.3 Business Analysis Methodology

The business analysis process took place in two phases. First, an adjustment of a
subset of the Tropos methodology [6] was used to extract business opportunities and
potential revenue streams. This phase included five distinct steps: (i) market analysis
that identified the existing stakeholders and business models related to the
investigated domains, (ii) scenario analysis that processed the final scenarios
produced for the technical requirements extraction in order to identify the related
stakeholders, the potential business interests that arise and the business opportunities
that emerge, (iii) generalization of the identified stakeholders that led to the
identification of the existing and new stakeholders that are involved in the envisaged
system, (iv) extraction of business requirements that are stakeholder specific and (v)
extraction of the business opportunities and the respective value proposition. Once the
process above was complete, the Business Model Canvas methodology [7] was
exploited to assist in defining the applicable business models. Thus, the business
model canvas approach has been used over the Discover, Connect and Organise
phases that contribute to the formation of the envisaged system. This resulted in
identifying how this system can offer value to various stakeholders; portray the
capabilities and partners required for creating, marketing, and delivering this value,
with the goal of generating profitable and sustainable revenue streams.
5 Architecture
The architecture that implements the concepts presented above is illustrated in Figure 4,
where an overview of the “core services” provided by the proposed architecture is
provided. The services depicted are grouped according to the major concept they
manipulate or operate on. Thus, services that operate on a single CIS are grouped
together, as are those that operate on a CSS, and those found on every node in a CSS.
Fig. 4. The Cooperating Smart Space layered functional architecture
Multi CSS/CIS Services operate for the benefit of more than one CIS, or more than
one CSS. Thus, they operate for a wider group of stakeholders. They offer federated
search and domain administration functions and require multiple CSSs or CISs to be
effective. This group includes the following services: the Domain Authority (that
provides and manages the CSS and CIS identities in a decentralised manner, allowing
authentication between multiple domains), the CIS Directory (that manages the CIS
information in a decentralised repository, it records available CISs within a domain or
set of domains, it enables searching for CISs based on specific criteria and it allows a
CIS to be removed from the repository), the CSS Directory (that provides search
38 K. Doolin et al.
facilities for CSSs, based on their identifier or by specifying search criteria, such as,
public profile attributes and tags), the CIS Recommendation (that is responsible for
handling CIS recommendations, allowing for recommendations of CISs to users and
vice versa, considering, among others, the users’ privacy preferences) and the Service
Market Place (that provides access to a repository of installable 3rd party (3P) services
and optional “core” services and provide mechanisms for accounting and charging).
CIS Services operate on behalf of a single CIS. There is at least one instance of
these services per CIS and an instance of these services can be used by multiple CISs.
The CIS services are: the CIS Management (that is responsible for handling all
aspects of CIS lifecycle management (creation, update and removal), provides control
over CIS membership and includes a community profile manager and a role manager
to specify the governance model for the CIS), the Community Context Management
(that enables access to and maintenance of community context, providing query
capabilities, as well as, addition/update/removal operations for community context,
maintaining the history of context for a CIS, and inferring community context
information), the Community Learning (that supports community preferences and
community intent learning) and the Community Personalisation (that manages the
community preferences and community intent and exposes interfaces for community
members to retrieve these preferences and intent models for their own use).
CSS Services operate on behalf of a single participant or CSS. There is at least one
instance of these services per participant and an instance of these services can be used
by multiple participants. The CSS services are: the CSS Management (that controls
which Nodes (devices or cloud instances) are part of the CSS, assigns a common
identifier and manages resource sharing & configuration policies), the User Context
Management (that is responsible for acquiring the user context from sensors and other
context sources, for modelling and managing the collected data, for maintaining
current & historic context in appropriate data repositories and for the provision of
inference techniques enabling the extraction of high level information from raw
context data), the User Personalisation (that manages & evaluates the user
behavioural models, such as user preferences, user intent, Bayesian models, etc., and
eventually identifies the actions that need to be taken), the Social Network Connection
(that integrates with existing Social Networking Systems (SNSs), enabling the
extraction of public info available in SNSs, as well as access/update of non-public
information for the specified user), the Privacy Protection (that provides identity
management mechanisms, facilities for managing the CSS privacy policies, which
specify the terms and conditions the CSS will respect concerning the personal data,
also offering Privacy Policy Negotiation facilities), the User Learning (that supports
learning of user behaviour models exploiting the user’s history of actions stored in the
system), the User Agent (that acts on behalf of a single CSS based on information
from several CSS and CIS components, aiming to establish the system’s proactive
behavior, resolving any conflicts that may arise, also enabling CSS users to provide
feedback on the system actions or decisions), the Trust Management (that is
responsible for collecting, maintaining and managing all information required for
assessing the trust relationships and includes a Trust Engine for evaluating direct,
indirect and user perceived trust) and the Service Provisioning (that supports the setup
and lifecycle control of a 3P service or CSS resource, allowing for installation,
(re)configuration and removal of new 3P services, also supporting the enforcement of
3P service sharing policies).
Node Services are available per CSS Node. A CSS Node is a logical node (device
or cloud instance) running CSS software that coordinates with other CSS Nodes to
form a participant’s CSS. There is an instance of these services per CSS Node. This
grouping includes the following services: the Communication Framework (that
provides the necessary mechanisms to support intra- and inter-CSS communication,
supporting the identification and maintenance of network connections, the discovery
of CSS Nodes (devices), and the communication between discovered nodes), the
Device Management (that provides mechanisms for managing devices within a CSS,
supporting the discovery of hardware devices and management of their capabilities)
and the Service Discovery (that provides service discovery and advertisement
mechanisms, enabling the discovery of core platform services within a CSS, as well
as, the discovery of 3P services shared by other CSSs or CISs).
6 Initial Evaluation
Using Paper Trials, an initial user evaluation was conducted in April 2011 across all
three user communities, i.e., the Disaster Management, the Student and the Enterprise
community. The primary objective of these trials was to record users’ responses to
early prototypes of initial scenarios & concepts and how users’ experiences of these
prototypes conformed to the previously identified user requirements. These Paper
Trials were interpreted loosely as a user evaluation trial of low-fidelity prototypes. A
secondary objective was to engage with users to confirm or discover the opportunity
spaces for pervasive and social computing, “where there is no urgent problem to be
solved, but much potential to augment and enhance practice in new ways” [8].
The envisaged system posed a challenge that could not be served by traditional
paper prototyping alone, since it required prototypes for user evaluation that focused
on user activities, goals and contexts of use, with varied levels of detail, thus
conveying a range of CSS/CIS system interactions within the user domains, which
were not necessarily focused on users manipulating device interfaces (i.e., pervasive
services working in the background for the benefit of their users). Therefore a specific
evaluation methodology was necessary.
6.1 Evaluation Methodology

Two user evaluation methods were selected for the trials, namely Storyboards
and Wizard of Oz. Both methods are known to be well suited to work with the
scenario-based design approach method, which was employed to describe potential
deployment for each of the three user communities. Both are methods that employ
scenario based vision prototypes [9], which serve the purpose of defining early design
focus for developers and providing a site for evaluating user responses [10].
Storyboards were the primary method used and acted as an informal low-fidelity
design artefact that provided a concrete but flexible example of how services and
features of the envisaged system, in this case an intelligent pervasive communities
system deployed in a context familiar to each user group, would be envisioned by the
consortium. The users’ feedback to our storyboards was captured using two different
techniques: (i) questions that were embedded in each storyboard and were addressed
to individual users in the context of viewing the storyboard, to focus their attention
40 K. Doolin et al.
and allow them to directly answer questions of particular interest to the project
researchers and (ii) participatory discussions that were facilitated after the storyboard
viewings, in the case of the Enterprise and Disaster Management communities, where
the storyboards provided springboards to openly discuss users’ reactions to the issues
and scenarios depicted within.
The Wizard of Oz method was used as a secondary method in the case of the
Student community. This method utilized a script based on university scenario
segments, with the project researchers playing the role of the envisioned system by
managing environmental and device responses to user activities and preferences in a
pervasive laboratory, which had been set up to stage an intelligent campus
environment. Participants answered questions posed during the experiment that was
also videotaped. It was designed to allow students evaluate an immersive experience
of a social and pervasive environment.
6.2 Evaluation Results

The overall feedback from this evaluation study did indicate strong support for the
concepts that were presented, albeit with quite a number of concerns expressed,
including: trusting the system, controlling privacy and difficulties with accepting
automation. It was evident that users did see the value concept of creating purpose-
driven communities and the leveraging of collective intelligence from those
communities, which is at the core of the project’s value proposition.
Regarding privacy and information disclosure, most students were happy to
disclose basic information. Most users were in favour of sharing their preferences and
requirements, although they were more reluctant to disclose other information and
wanted to be consulted on disclosures rather than letting their devices make decisions
for them. This was part of a general concern regarding trusting a system to make
automatic decisions on behalf of users.
There was a common concern between the student and enterprise users on how far
technology should go in replacing natural human behaviours. The students considered it
acceptable if the system suggests greeting somebody based on shared interests or
intents, as well as the person’s mood, but only a few students would use this to start
friendships, preferring to meet new people “the old-fashioned way” instead. The
enterprise users liked the professional networking support features, but some questioned
the effect it may have on the natural networking activity and suggested it could rule out
opportunistic encounters. One user stated: “A user could become a slave to their
preferences and may lose the opportunity to discover new opportunities”.
It is clear that a main objective for the project’s researchers is to design a system
for creating purpose-driven communities, through rich context data sets. Yet, it is also
clear from our user responses that a key concern is that the system appears to be too
intrusive and that there would be serious concerns around the areas of privacy, trust,
automated community creation and service delivery. This appears to be a paradox,
since the system requires the user to give rich context data in order to provide context-
aware personalised services and as a result is a key challenge for the consortium.
7 Conclusions
The SOCIETIES project aims to investigate and address the gap between pervasive
and social computing by designing, implementing and evaluating an open scalable
service architecture and platform. Based on a vision of the discovery, connection and
organisation of relevant people, resources and things into dynamically formed
pervasive communities, SOCIETIES attempts to bridge the domains of pervasive and
social computing in a unified platform allowing individuals to utilise pervasive
services in a community sphere.
This paper presents concepts and research methodologies adopted in the
SOCIETIES project towards the realization of Pervasive Communities and in order to
assess whether real end users can see a value for engaging with such a system. The
overall feedback from our initial user evaluation study did indicate strong support for
the concepts that were presented, albeit with quite a number of concerns expressed,
including: trusting the system, controlling privacy and accepting automated decision
making. Based on these user concerns and considering every user feedback collected
from the initial trials, the technical requirements have been revised and the
SOCIETIES architecture has been adapted accordingly. Two more user trials have
been scheduled for 2012 and 2013 that will enable us to assess how successfully the
vision, concepts and results of SOCIETIES address the technological and user
acceptance gap between pervasive and social computing. The results achieved up to
this point and the user feedback already collected indicate that the SOCIETIES
platform can find its way to the facilities portfolio that users exploit on a daily basis.
References
1. Saha, D., Mukherjee, A.: Pervasive computing: A paradigm for the 21st century. IEEE
Computer 36(3), 25–31 (2003)
2. Dasgupta, S.: Social Computing: Concepts, Methodologies, Tools, and Applications. IGI
Global (2010)
3. Muller, M.J.: Participatory Design: The third space in HCI. In: The Human-Computer
Interaction Handbook: Fundamentals, Evolving Technologies and Emerging Applications,
pp. 1051–1068. Lawrence Erlbaum Associates, New Jersey (2002)
4. Wright, P., McCarthy, J.: Empathy and Experience in HCI. In: 26th Annual SIGCHI
Conference on Human Factors in Computing Systems, Florence, Italy (April 2008)
5. Grady, R.B., Caswell, D.L.: Software Metrics: Establishing a Company-Wide Program.
Prentice Hall, Englewood Cliffs (1987)
6. Castro, J., Kolp, M., Mylopoulos, J.: Towards Requirements-Driven Information Systems
Engineering: The Tropos Project. Information Systems 27(6), 365–389 (2002)
7. Osterwalder, A., Pigneur, Y.: Business Model Generation: A Handbook for Visionaries,
Game Changers, and Challengers. John Wiley & Sons, New Jersey (2010)
8. Hornecker, E., Halloran, J., Fitzpatrick, G., Weal, M., Millard, D., Michaelides, D.,
Cruickshank, D., Roure, D.D.: UbiComp in opportunity spaces: challenges for
participatory design. In: 9th Conference on Participatory Design, Trento, Italy (August
2006)
9. Erickson, T.: Notes on Design Practice: Stories and Prototypes as Catalysts for
Communication. In: Scenario-Based Design: Envisioning Work and Technology in System
Development. John Wiley & Sons, New York (1995)
10. Houde, S., Hill, C.: What do prototypes prototype? In: Handbook of Human-Computer
Interaction, 2nd edn., Elsevier Science B. V., Amsterdam (1997)
Cross-Disciplinary Lessons for the Future Internet
Anne-Marie Oostveen1, Isis Hjorth1, Brian Pickering2, Michael Boniface2,

Eric T. Meyer1, Cristobal Cobo1, and Ralph Schroeder1
1
Oxford Internet Institute, University of Oxford, United Kingdom
{anne-marie.oostveen,isis.hjorth,eric.meyer,cristbal.cobo,
ralph.schroeder}@oii.ox.ac.uk
2
University of Southampton IT Innovation, United Kingdom
{jbp,mjb}@it-innovation.soton.ac.uk
Abstract. There are many societal concerns that emerge as a consequence of

Future Internet (FI) research and development. A survey identified six key
social and economic issues deemed most relevant to European FI projects.
During a SESERV-organized workshop, experts in Future Internet technology
engaged with social scientists (including economists), policy experts and other
stakeholders in analyzing the socio-economic barriers and challenges that affect
the Future Internet, and conversely, how the Future Internet will affect society,
government, and business. The workshop aimed to bridge the gap between
those who study and those who build the Internet. This chapter describes the
socio-economic barriers seen by the community itself related to the Future
Internet and suggests their resolution, as well as investigating how relevant the
EU Digital Agenda is to Future Internet technologists.
Keywords: Future Internet, Socio-Economics, Digital Agenda, Users, SESERV.
1 Introduction
The Internet has become an essential part of the infrastructure of modern life.
Relationships are managed online, commerce increasingly takes place online, media
content has moved online, television and entertainment are being delivered via the
Internet, and policy makers engage the public via programs such as Digital Britain
[1], the European Digital Agenda [2], and other worldwide initiatives. Efforts to
develop the so-called Future Internet (FI), will either follow as a logical extension of
what is in place now, or as something completely different [3].
At the same time the Internet’s underlying technology is evolving, it is also
changing as a social and economic platform. Yet it is not clear how competing
interests should be balanced when technical, societal, economic and regulatory
concerns come into conflict. One view is that technology developers should develop
innovative technologies with little oversight and regulation so as not to stifle
creativity. Social and regulatory concerns can be dealt with as they arise as a result of
use. A user-centric view, on the other hand, suggests that any FI must be designed
around social and economic concerns, with technology that supports values such as
inclusion, privacy, and democracy.
Cross-Disciplinary Lessons for the Future Internet 43
Innovation is often serendipitous [4]; for maximum benefit, the complex

interactions and even antagonisms between society and technologists need to be
nurtured in a suitable and enabling environment. Thus social, legal and technical
perspectives inevitably intertwine. Understanding the interactions between
technologists, society, legislation and regulation is therefore indispensable in shaping
the Future Internet and associated applications and services [5, 6]. In this chapter we
investigate the societal aspects of the FI as seen by social scientists, policy makers
and technologists involved in the central European Commission-funded projects
designing these technologies.
How the Internet pervades our professional, commercial, political and leisure
activities is an important question for Europe and beyond. Boosting EU R&D efforts
is a key element of the Digital Agenda for Europe [2]. EU-funded research aims to
make the Internet of the future a dynamic place for innovation, growth and jobs. The
European Commission is currently reviewing the progress of some 140 "Future
Internet" research projects which it supports. Given the relevance of planned Digital
Agenda actions for the SESERV workshop’s participants and their proximity to
several themes included in the programme, it seemed important to learn how familiar
with this EU instrument they are and the value it provides to their current activities.
Therefore ten participants were interviewed on this topic.
The specific socio-economic topics discussed during the workshop ‘The Future
Internet: The Social Nature of Technical Choices’ organized by the SESERV
consortium1 were based on the results of an online survey across the FI community. The
structure of this chapter is as follows: in Section 2 we discuss the socio-economic topics
that emerged via representatives of FI projects as they relate to any barriers they face in
their development work. From such discussions, eight cross-cutting strategies emerged
that provide potential resolutions to these socio-economic challenges (Section 3).
Finally, in Section 4 we identify how relevant the Digital Agenda is to Future Internet
technologists and examine its value for the projects interviewed.
2 Societal Concerns and Challenges

In 2010, the Internet Society defined an ‘Internet Ecosystem’ [7], with stakeholders
from a traditional infrastructure perspective. In recent years, however, the rapid
convergence of technologies has increased the scope of stakeholder engagement
beyond what was originally described. The European FI initiative has led
developments both within the core ICT programme and the Future Internet Public
Private Partnership (FI-PPP) initiative2. A significant increase in the diversity of roles
is seen, along with an increased emphasis on users in addition to infrastructure and a
blurring of roles between major market players [8]. The concerns of the Internet have
moved from structures for the delivery of data, to socio-economic structures
supporting information and knowledge exchange.
Many societal concerns emerge as a consequence of FI research and development.
Relating these specifically to the FI ecosystem rather than to more general societal
1
SESERV (Socio-Economic Services for European Research Projects). See http://www.
seserv.org
2
The Future Internet – Public Private Partnership, http://www.fi-ppp.eu/
44 A. Oostveen et al.
issues is essential FI technology projects in debate. Content analysis of two recent

reports, Social Impact of ICT Studies [9] and Towards a Future Internet [10]
identified 16 societal concerns for the FI that raise significant technical, commercial
and regulatory challenges: (1) Regulation; (2) Privacy; (3) Online Identity; (4) Green
Issues; (5) Security of Communications; (6) Content Regulation; (7) Cloud
Computing; (8) Trust; (9) e-Democracy; (10) Digital Citizenship; (11) Digital
Inclusion; (12) Online Communities; (13) Internet of Things; (14) Consumers and
Suppliers; (15) Distributed Knowledge; (16) Cybercrime and Cyberlaw.
Representatives from FP7 Future Internet projects (n=98) rated the relevance of
these socio-economic topics for their projects on a subjective scale from “Not
Relevant” through to “Absolutely relevant, a key issue” in an online survey. The
following six issues were of most interest: Privacy and Data Protection including
user data, file-sharing control, selling of personal information; Online Identity
including anonymity, digital presence, rights to delete information; Security of
Communications including legal implications; Online Communities including social
networks, virtual relationships; Internet of Things and the connections between people
and devices; and Cloud Computing including the risks and benefits of virtual access to
information. Some topics (Green Internet and Cybercrime, as well as Digital
Inclusion) were disregarded by all projects, while applied to only a few [11].
During a workshop and seminar held at the University of Oxford in June 2011,
experts in FI technology engaged with researchers such as social scientists (including
economists), policy experts and other stakeholders to explore the socio-economic
aspects of the FI, and conversely how the FI will affect society, government, and
business [12]. Special break-out sessions on each of the six key issues were organized
to facilitate a more focused discussion between the 69 participants, with the societal
concerns and challenges from these 1.5 hour break-out sessions discussed below.
2.1 Privacy and Data Protection

As the Internet becomes more integral to the way we live our daily lives, end users are
becoming increasingly aware of the dangers of making too much information
available publicly [13]. Careers and personal lives can be severely affected by not
considering what information (including multimedia – photos, videos etc.) is
disclosed online. For most users, the main concern is the extent to which information
was becoming public, and some are now allowing less of their content to be published
openly. This change in general awareness will make FI applications safer (e.g.,
customers and regulators will demand that location-aware services protect user
privacy). But while attitudes towards privacy are changing significantly, for many the
level of privacy concern is decreasing.
Privacy is heavily compromised by a lack of awareness as much as by technical or
cost issues. Users supply personal information to service providers with every post,
query or click in applications like Google Search, Facebook, and Twitter. Users
benefit from this data exchange because they can use search technology, social
networks and the like without charge. Yet the relationship between citizens and
service providers is highly asymmetric, and the resulting loss of privacy for users and
bystanders is profound. The providers of these services exploit this content in a wide
variety of ways: to attract a larger audience share; to classify users based on their
personal data to ‘improve’ the service; to classify and index data (including personal
relationship data) which allows the service to be further enhanced; to create
personalized advertising; and to provide information to businesses and governments,
for payment and/or to meet legal obligations.
The most successful Social Network Sites or online retailers are now among the
largest and most profitable businesses, and yet typically accept no responsibility for
3
user-generated content . Users can publish sensitive, sometimes scandalous
information about third parties, which is propagated freely by the service provider.
The victims have few protections and very limited recourse. They can ask the service
provider to remove the offending content after the fact, or sue the user who posted it
(if the service provider reveals their real identity, and that user falls under a
jurisdiction to which the victim has access).
The trend is towards an increase in asymmetry as service providers improve
exploitation and find new opportunities to capture personal data. Personal data is
increasingly available to the service provider and to other users, commercial customers
and government agencies. The risks from widespread disclosure - should the provider
be hacked or forced by government agencies to release information - are acute.
European privacy regulations provide little protection due to technical and jurisdictional
limitations; European service providers may therefore find it harder to compete.
Privacy clearly goes hand-in-hand with issues of security and trust. Therefore, one
could expect appropriate technical and procedural protection in support of users
online. To some degree, users may have unrealistic expectations of technical
provision for privacy. However, it is equally true that users themselves should be able
to make appropriate judgments about suitable protection and data management. Thus,
examining how users behave and wish to behave may help determine requirements.
2.2 Online Identity

Online identity is inextricably related to issues of data, privacy and rights (including,
though not limited to, digital rights). The concern today has switched to the more
fundamental question of how identity is to be understood within the context of (user)
interactions in different socio-technical environments. It thus becomes necessary to
examine the relationships between all data and identity.
Identity is not easy to define, and current definitions diverge. Common baselines
and vocabularies are needed to enable a multidisciplinary discussion of identity.
Society conceives identity as stable: identity in terms such as surname and passport
and the like is assumed stable by policy-makers and in terms of social norms. Yet, in
scholarly discourses and research on identity, it is often characterized as inherently
dynamic (changing over time and context). In addition, individuals might very well
experience their identity as fluid or develop multiple identities [14]. This clash
between these two opposing stances is not sufficiently addressed.
A number of socio-technical challenges arise. First, there is a need to develop tools
for managing online identity. As applications are increasingly tied to each other, users
need assistance in understanding the implications of these connections for the sharing
of their data and identity/-ies. Designing tools that enable multi-scale filtering of
3
Though this is not always the case, e.g. Italian law puts the onus on the service provider.
content by users (e.g. more control of what information is accessible to whom) is an

immediate challenge to be addressed.
Second, in an online/networked environment, users leave digital footprints. These
data can be misused by third parties. In addition, more sophisticated methods for
analyzing large-scale data from, for example, archived system logs, mobile phone
usage, and other online interactions make it possible to identify individuals based on
their preferences, patterns and social networks. Sometimes it’s justified (mobile
phone usage for billing), but generally anonymization is desirable. This places an
increased onus on developers, legislators, third parties and researchers to disclose the
degree to which data reveal identity.
Third, currently anonymity cannot be guaranteed online and individual users can,
with some effort, almost always be identified. Users need to know the levels of
anonymity possible. This leads on to the question whether anonymity should form
part of a more general set of digital rights. One challenge then is to develop features
that allow for increasing levels of transparency: end-users could be made aware of the
level, or lack of, anonymity that systems allow for.
Finally, the right of an individual ‘to be forgotten’ poses specific problems. This
relates directly to the interplay between an individual’s rights and those of the
community. Are there occasions so significant or horrific an individual’s identity
online should not be protected, in the interests of the common good?
2.3 Security of Communications

Security of communications is not about privacy or identity management. Instead, it is
about managing the risks to the smooth functioning of critical and non-critical
infrastructures, to financial stability, and to personal security and trust. Security in this
context, therefore, is about risk management.
Cloud computing is a fundamental component within the FI ecosystem. While
cloud computing could provide access to vast resources, clouds raise concerns about
the risks they pose. For instance, what if cloud providers or their customers were
malicious? If we cannot protect the data, how can we guarantee that the services can
be protected? Who should be responsible for meeting the security threats of clouds:
the operator, developer or customer, or even the regulator? One extreme scenario
could be that the cloud provider becomes the key party responsible for the cloud with
worrying implications for the degree of freedom of users. In contrast, little or no
regulation could be a risk to parts of the innovation, as a deterrent to creative FI
services. And any legislation needs to be cross-jurisdictional.
Even when compliant to existing EU legislation concerning storage and privacy,
the nature of the cloud brings new risks. Many SMEs are thinking of moving their
regular ICT needs into a cloud and for a smaller company, it could be better not to
impose regulation, especially if it lags behind innovations. Service providers could be
compelled to manage the risks, and customers need to trust the infrastructure
provider. But over-monitoring may make users distrust the service.
Security can be addressed via technical requirements, but the more difficult emerging
challenges are socio-economic: what are the obligations of those who did not expect to
be supporting these services? Access to risk expertise and managing risk are essential. A
cloud provider has a team of security analysts or information security analysts, and large
corporations employ legal services firms. Others, however, may not have access to risk
experts or be able to cope with security threats. Most medium and small scale
companies cannot afford to hire technical risk analysts, lawyers and other experts.
Similarly, domestic users will have to trust the information provided. Security could be
left to the market, with customers avoiding services that they find too risky. But the
laissez-faire of a completely free market is not enough to manage security risks. There
is a need for regulation, and one simple approach could be to force cloud service
providers to publish statistics about the health of their activities and their monthly
attacks, allowing for validation. Yet information about security is also very sensitive,
which means that service providers might not be willing to reveal these data. Hence
there is a need for transparent metrics for comparing ‘trustworthiness’ and auditing
standards to ensure that what service providers publish is credible.
2.4 Internet of Things

Definitions of the Internet of Things (IoT) vary. At a minimum, the IoT can be
thought of as including all manner of mobile devices, including telephones, PDAs and
sensors equipped with intelligent and large-scale data analytics. The key ingredient is
the seamless interaction between different systems: IoT technologies are bringing data
together to create new services. The promise of the IoT is to use online technology
combined with sensors which might automate the surveillance and management of the
more mundane aspects of life (food purchases which are linked to fridge monitors;
automation in the home; and so forth).
Many barriers have been identified for the adoption of the IoT within the FI
ecosystem. First, participants indicated that current definitions are too abstract and
hard to grasp, too academic without enough focus on design and applications. This is
partly due to the lack of interaction between the actors in the design and application
domains. Currently, development is characterized by 'doing' rather than by reflexivity
and deliberations about design. Even so the general public perceives the IoT in terms
of Big Brother: ‘Smart’ applications tend to be received with skepticism by the
general public, such as the ‘smart’ bins in London provided with sensors which were
quickly labelled ‘spy’-bins [15]. In popular discourse, technologies are described as
intelligent autonomous agents ‘affecting’ a passive public. Changing this attitude and
the underlying technologically deterministic view would help to inform design.
IoT technologies are predominantly designed for domestic purposes, such as the
interactive ‘intelligent’ Internet fridge. Applications need introduction in existing
infrastructures such as transport and health systems to make them more intelligent.
Additional challenges are the vast amounts of data generated. Individual systems,
however, are not able to harness the data and so we need an ‘intermediate’ level of
technology4. Further, where are boundaries between public and private data? One
example is the ‘passive’ monitoring phones: with mobiles on, users can be tracked at
all times. As well as transparency, the advantages and disadvantages (e.g. spam risks)
need to be weighed up. Users could, for example, be presented with different levels of
4
Possibly by extending senslets, http://www.inets.rwth-aachen.de/fileadmin/
templates/images/PublicationPdfs/2008/Senslet_EuroSSC.pdf
‘sign-off’ options to balance against the possibility of generating moral panics by

greater awareness. It is also vital to provide opportunities for 'offline' access to
services; ‘opting out’ currently unacceptably penalizes people.
Finally, as ever, there may be unintended consequences. An example from the
health sector: Some elderly people have sensors implemented in their homes,
measuring levels of moisture. While such sensors can help alert carers, they might
also see human expertise replaced by automated sensors. Such effects are important.
2.5 Online Communities
Social media have grown rapidly – today nearly 4 out of 5 active internet users visit
social networks and blogs [16]; 20% of online time is spent on social networking sites
(SNS’s), up from 6% in 2007. SNS’s reach 82% of the world’s online population
[17]. Online communities center on how users interact with and exploit the range of
social networking applications (e.g., government, leisure and work). A critical success
factor is to maximize activity, mainly achieved irrespective of the purpose of
communications. However, it is also necessary to comply with required data
protection legislation in relation to responsibilities and individual actions (e.g.
consent). Herein lies a contradiction: Privacy compliance, often promoted as a means
to increase trust and hence participation, can also act as an inhibiter to greater activity.
Individuals use SNS’s because their perception of risk is considered low enough,
whilst developing an appetite for risk, upping participation regardless of associated
regulation.
This leads to an interesting challenge for European service providers and research
projects: How to strike the balance between participation and privacy - if it is
desirable to monitor and mine data - without violating a citizen’s right to privacy. It is
unlikely that the successful paradigms of the last decade, social networking and
clouds, would have prospered if they’d been subject to the European regulatory
environment from the start. The try-it-and-see approach has led to a balance over
time: participants have explored their preferences iteratively. Social networking has in
fact been a large experiment in people’s appetite for privacy.
Online Communities highlight the basic dichotomy: is it technology or society
which shapes the ICT future? The answer for now at least is that there is a real need to
back off from technology for technology’s sake and begin to take seriously how
communities are formed and what they do online. The focus would move towards
societal behaviours and away from technology, and require appropriately skilled
cross-disciplinary researchers with an understanding of these communities and what
makes healthy and vibrant online communities.
Elsewhere, SNS content (especially user profiles) are being synchronized live
across networks. What does this do for user control and user-centeredness? User-
centric platform-bridging applications with transparent filtering options can be
developed, so users should be able to manage and control sharing easily with the
online communities. Better tools in general are needed for managing online
communities such as smaller community hubs that mirror the cognitive limit for
social relationships. There are both limitations and strengths to smaller online
communities: there is less information accessible but smaller communities could be

one way of handling privacy issues and the right to be forgotten (see above) in line
with community benefit.
Finally, users make innovative and creative use of systems and applications in the
development of online communities. Technologies are not the only drivers in the
development of new types of online communities where different structures may be
required for sharing or co-creating content. There is a need to balance bottom-up and
top-down technology development, and to involve members of the communities.
2.6 Cloud Computing

Just as energy production benefits from economies of scale when consumers transfer
responsibility to an electrical grid for centralized production, so do those needing ICT
resources benefit from exploiting cloud facilities. Europe could gain significantly
from the resulting new business opportunities even though it lags behind the rest of
the world with clouds, not least because much of European enterprise is SME based
for whom investment in large and under-used ICT equipment may not be economic.
Early end-user engagement is critical to direct investment and design. At the same
time, of course, issues of trust and security cannot be overlooked and these need to be
tackled alongside interoperability and portability.
There are a number of barriers to the adoption of cloud computing within the FI
ecosystem, such as the lack of a global legal framework. The global nature of cloud
computing requires consistency in laws across jurisdictions (e.g. to notify data access
breaches). International coordination is important here but also bottom-up feedback
from users. Definitions also pose problems with clouds: are they infrastructure or do
they encompass nearly all online activity? Another barrier is that EU discourse
focuses on risks and less on benefits, especially economic ones, and is slow to adopt
new technology, sticking for instance with grids instead.
User concerns relate largely to control. There is a need for more transparency and
control. Contracts vary greatly between different providers and often do not allow
user control over where their data is stored; many companies run services on a third
company's cloud infrastructure; end-users don't deal directly with the cloud provider
and yet rely upon them to secure the data and provide the actual service. Security in
general is a concern, though is tightly coupled with transparency. Designing for
interoperability and portability while allowing customization is also of concern.
Portability will allow users to move from one cloud provider to another and avoid
platform lock-in. Finally, providers might gain a large amount of meta-data about the
activities, locations, and contents of user interactions with their services; again
transparency would be appreciated.
3 Cross-Cutting Resolutions to Socio-Economic Challenges
The discussions in Section 2 yielded recurring strategies which suggest eight cross-
cutting resolutions to the socio-economic challenges identified.
3.1 Call for Increased Transparency

A dominant trend across discussions was a call for increased transparency on all
levels for end-users of networked ICTs. Systems and applications should offer end-
users tools that allow end-users to know exactly who has access to the contents of
their online activities. Advanced transparent filtering options are becoming
increasingly critical as more and more online networks are being synchronized, as are
tools that assist users to manage the various communities.
Transparency also relates to ISPs and data storage, particularly with cloud-based
services. To make security risks more transparent for end-users, providers might
publish monthly statistics on attacks. End-users should be able to easily identify
where and how their data is stored and is or will be used.
3.2 Call for More User-Centricity and Control

Discussions converged on a call for more user-centricity and control: increased user-
centricity in the design of applications. Users could be allowed some means of
influencing applications/systems on an ongoing basis; creative uses could feed back
into systems to improve them and innovate further. Control is particularly evident in
the context of opt-out options with more granularity required. Additionally, a range of
different choices for how user data is stored could be offered (e.g. location). Finally,
users need to assess and control their security risks and risk management.
3.3 Continuing Need for Further Multi-disciplinary Bridging

Without exception the discussions called for increase cooperation across sectors.
While it is easy to call for knowledge-exchange, dialogue and collaboration across
and beyond academic fields, industry, developers, designers and users gaps exist
between privacy researchers and IoT engineers, or between eHealth practitioners and
IT suppliers, for instance. Creating frameworks for knowledge exchange between
users, developers, regulators and researchers would facilitate connection between
technical and legal analysts and a better understanding of risks could avoid
‘siloization’ or ‘pillarization’. The expertise of different communities should be
included in all stages of technology development and design via multi-disciplinary
engagement and institutions.
3.4 Striking a Balance between Extremes in Debates and Design
A cross-cutting theme that emerged across several discussions (Online Identity and
Communities, the IoT, and Privacy) was a call for more balanced approaches in
design avoiding dichotomized thinking. For example, there is a need for a balance
between identity as singular and stable (e.g. passport) as well as completely fluid and
dynamic. How identity is perceived has a consequence for system design such as
more nuanced views and multi-disciplinary insights, like an identity continuum from
stable to dynamic. Similarly, design needs to balance bottom-up and top-down
innovation: new forms of communities are potential drivers of technology

development. Elsewhere, eHealth privacy practices and perceptions suggest another
balance to strike: a middle ground that allows proportionate access to patient records
rather than either a laissez-faire approach or over-regulation would be beneficial.
Finally, discourses on privacy tend to lack balance between risk and opportunity: the
IoT technologies, are often perceived as 'big brother' surveillance, for instance.
3.5 Facilitating the Development of Digital Literacy

The need for greater digital and media literacy education was expressed across
sessions (Security, Privacy, Identity and Online Communities) the core concerns
being user ability to critically manage privacy and identity. Arguably, digital literacy
skills can equip users with more sophisticated tools for managing and understanding
identity and thus solve some of the problems encountered with privacy. Security risks
could be managed better with best practice guidelines and more awareness. This
highlights non-technical social challenges that need to be addressed alongside the
design and development of socio-technical systems.
3.6 Addressing the Lack of Common Vocabularies and Definitions
Common vocabularies and better definitions (Identity; Internet of Things; Online

Communities; Cloud Computing) have the potential to be enablers: in cloud
computing current definitions diverge between infrastructure and all online activities.
For the IoT definitions are too academic, lack focus on design, and difficult to apply
in technology development. For identity, there is a need for definitions that
acknowledge a close link with questions of privacy, data and rights in digital contexts.
Common vocabularies could benefit new technologies and their adoption. For now,
they are missing, in the case of the multi-device IoT. Likewise, a more advanced
vocabulary is needed to describe the maintenance, structure, and scales of online
communities. Seen in light of multi-disciplinary bridging and collaboration, there is a
need for adequate vocabulary and definitions that can be applied across sectors and
contexts.
3.7 Need for Clarity about Digital Rights and Digital Choice
Some discussions (Privacy, Internet of Things and Online Communities) agreed on
the need to clarify digital rights and digital choices: what levels of anonymity should
be granted, to whom and in what context? In the case of eHealth, for example, there is
a need to balance an individual's right to anonymity against appropriate access to
detect and tackle emerging health issues. Another question concerns the right to be
forgotten: to have information deleted. As stated, this might not apply to content of
historic or humanitarian value. Digital choice can be exemplified in relation to the
IoT, where off-line alternatives should be available.
3.8 Enabling Global Regulatory Frameworks

Global regulatory frameworks are particularly pertinent (Security, Online
Communities and Cloud Computing). Suggestions here include consistency across
jurisdictions for data breaches as well as for anonymity. Increased trans-national
legislation could ensure that providers are not discouraged from operating in certain
countries (e.g. where providers are liable for users’ IP infringements).
4 The Future Internet Community and the Digital Agenda

ICT is regarded as increasingly critical for the future growth and development of
Europe. Europe 2020 [18] together with the Digital Agenda [2] outlines the main
challenges and opportunities over the coming decade including for the FI. The overall
aim of the Digital Agenda is to “deliver sustainable and social benefits from a digital
single market based on fast and ultra-fast internet and interoperable applications”.
At its center is an assumption about the mutual reinforcement between innovation
in the ICT sector and consumption which, in turn drives technological improvement.
This virtuous cycle runs something like this: if there are attractive services and
content available online across all member states this will motivate increased demand.
More users will want access, and look for more and improved content and services.
Increased demand in turn provides the necessary financial basis for improvements in
the supporting infrastructure. This investment enables ever more sophisticated service
and content generation and support, and so on.
Against this background, the Digital Agenda recognizes some seven major
challenges or obstacles: fragmented digital markets, lack of interoperability, rising
cybercrime and low trust, lack of investment in networks, insufficient R&D, lack of
skills, and fragmented answers to societal questions, which relate principally to
infrastructure and commerce; and the virtuous cycle must address these obstacles.
The previous sections have highlighted that the FI is of interest to different
stakeholders, and particularly the role of users in terms of improving technology
design and alleviating fears around privacy and security risks. These social aspects
should not be down-played in the Digital Agenda. The focus on infrastructure and
cross-border eCommerce fails to give a central place to end-users. The assumption of
the virtuous cycle is that end-users will participate. If so, considerable effort needs to
be invested in understanding the use of services and the inhibitors to online activity.
The Digital Agenda needs to engage closely with the FI community. Knowledge of
the aims and relevance of the Digital Agenda is highly variable across European ICT
projects and actors. A number of informal interviews with participants in this
community were conducted, and while perhaps not representative, clearly the projects
had little widespread understanding of the Digital Agenda’s aims. If familiar at all, it
was seen as irrelevant to the specific concerns within the projects themselves. Europe
may set an agenda and provide motivation for technology advance, but its relevance
and meaning for projects is unclear. Some believe the EU should not seek to micro-
manage projects: if innovation is to deliver, a large amount of autonomy is required.
Especially in discussions of the Internet of Things, designers and business developers
view the Digital Agenda as a restriction on new business plans and technology
designs. This also affects global competitiveness. Even so, there was a general
consensus that the Digital Agenda is central to taking Europe forward technologically
as well as socially: though too high-level lacking global relevance beyond the EU, as
an instrument for future strategy, technologists and social scientists have much to
contribute to the Digital Agenda and vice versa.
5 Conclusions
This chapter has presented the views of social scientists and technologists working on
the FI. The community has developed possible future strategies and priorities. The
results represent a snapshot of the challenges facing those undertaking FI research.
There is no doubt that the FI ecosystem is an increasingly rich, diverse and complex
environment, and Challenge 1 projects are aware of societal concerns and challenges,
and of their potential resolution. In contrast, the Digital Agenda is not well understood
by technologists and there is a gap between a set of high level policies and incentives
that are particularly focused on infrastructure and complex regulatory processes as
against the users of the technologies being developed. Regulations currently ignore
some of the concerns of citizens and there is a disconnect between the ‘stakeholders’
of the FI and the Digital Agenda. The European Commission needs to find a way to
update the Digital Agenda in response to the needs of a broad spectrum of people and
communities rather than focusing only on big companies or governments. For
instance, rural and remote regions, non-organized communities and even SMEs seem
to be under-represented in this policy aimed at 2020: different ‘soft’ design
mechanisms may help the Digital Agenda to adapt to the social, political, educational,
labour, and environmental needs of the community. If the Digital Agenda is not
embedded in the principles of openness, adaptability, participation and transparency,
it is hard to see how it will succeed. Supporting technologists in their understanding
of the potential broader impacts of the FI and its adoption through dialogue with
social scientists must be central to this effort. To realize the benefits for the widest
possible range of stakeholders, there will need to be increasing engagement between
those who study and those who are building the Future Internet.
Acknowledgements. The SESERV project (FP7-2010-ICT-258138-CSA) is funded
by the European Commission. The authors would like to thank all participants to the
Oxford scientific workshop and especially Ben Bashford, Ian Brown, Tony Fish,
Sandra Gonzalez-Bailon, Christopher Millard and Mike Surridge for facilitating.
References
1. Mandelson, P., Bradshaw, B.: Digital Britain: Final Report. Department for Business
Innovation & Skills and Department for Culture, Media and Sport (2009), http://
www.official-documents.gov.uk/document/cm76/7650/7650.pdf
2. European Commission: A Digital Agenda for Europe. European Commission (2010),
http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:
2010:0245:REV1:EN:HTML
3. The Internet Society: How is the Internet evolving? (2011), http://www.Internet

society.org/internet/how-it’s-evolving/future-scenarios
4. SESERV: The Future Internet Debate: Will the Design of the Future Internet be Driven by
Technology or Societal Concerns? (2011), http://www.seserv.org/panel/
conferences-webcasts#debate
5. Callon, M.: Society in the Making: The Study of Technology as a Tool for Social
Analysis. In: Bijker, W.E., Hughes, T.P., Pinch, T. (eds.) The Social Construction of
Technological Systems, pp. 83–103. The MIT Press, Cambridge (1987)
6. Hausheer, D., et al.: Future Internet Socio-Economics – Challenges and Perspectives (2009),
http://www.future-internet.eu/fileadmin/documents/madrid_
documents/FISE/FISE_position_paper_final.pdf
7. The Internet Society: The Internet Ecosystem (2010), http://www.isoc.org/
pubpolpillar/docs/internetmodel.pdf
8. Boniface, M., et al.: First Report on Economic Future Internet Coordination Activities.
SESERV Deliverable D2.1, Socio-Economic Services for European Research Projects
FP7-2010-ICT-258138-CSA (2011)
9. Universitat Siegen: Final Report on the Social Impact of ICTs in Europe to the European
Commission, SMART No2007/0068. European Commission (2010), http://ec.
europa.eu/information_society/eeurope/i2010/docs/eda/social_
impact_of_ict.pdf
10. Blackman, C., et al.: Towards a Future Internet: Interrelation between Technological,
Social and Economic Trends. European Commission (2010), http://www.
internetfutures.eu/wp-content/uploads/2010/11/TAFI-Final-
Report.pdf
11. Boniface, M., et al.: Initial SESERV Survey Results, Challenge 1 Projects: Socio-Economic
Priorities. SESERV (2011), http://www.seserv.org/fise-conversation/
socio-economicpriorities/SESERV-Survey-Results_May11.pdf
12. Oostveen, A., et al.: First Year Report on Scientific Workshop. SESERV Deliverable
D1.2, Socio-Economic Services for European Research Projects FP7-2010-ICT-258138-
CSA (2011), http://www.social-informatics.net/SESERV_D1.2.pdf
13. Pötzsch, S.: Privacy Awareness: A Means to Solve the Privacy Paradox? In: Matyáš, V.,
Fischer-Hübner, S., Cvrček, D., Švenda, P. (eds.) The Future of Identity. IFIP AICT,
14. Jones, S.R., McEwen, M.K.: A conceptual model of multiple dimensions of identity. J.
Coll. Student Dev. 41, 405–414 (2000)
15. Harris, E.: Smart wheelie bins to charge for waste. The London Evening Standard, London
(2007)
16. Nielsen: State of the Media: The Social Media Report Q3 2011 (2011), http://
blog.nielsen.com/nielsenwire/social/
17. Radwanik, S., et al.: It’s a Social World: Top 10 Need-to-Knows About Social Networking
and Where It’s Headed (2011), http://www.comscore.com/Press_Events/
Presentations_Whitepapers/2011/it_is_a_social_world_top_10_
need-to-knows_about_social_networking
18. European Commission: Europe 2020: A strategy for smart, sustainable and inclusive
growth. European Commission (2010), http://eur-lex.europa.eu/LexUri
Serv/LexUriServ.do?uri=COM:2010:2020:FIN:EN:PDF
Design Principles for the Future Internet Architecture
Dimitri Papadimitriou1, Theodore Zahariadis2, Pedro Martinez-Julia3,

Ioanna Papafili4, Vito Morreale5, Francesco Torelli5, Bernard Sales1,
and Piet Demeester6
1
Alcatel-Lucent Bell Labs, Belgium
{dimitri.papadimitriou,bernard.sales}@alcatel-lucent.com
2
Synelixis Solutions, Greece
[email protected]
3
Univ. of Murcia, Spain
[email protected]
4
Athens University of Economics and Business (AUEB)
[email protected]
5
Engineering, Italy
{vito.morreale,francesco.torelli}@eng.it
6
Ghent University (UGent), Belgium
[email protected]
Abstract. Design principles play a central role in the architecture of the Internet
as driving most engineering decisions at conception level and operational level.
This paper is based on the EC Future Internet Architecture (FIArch) Group
results and identifies some of the design principles that we expect to govern the
future architecture of the Internet. We believe that it may serve as a starting
point and comparison for most research and development projects that target the
so-called Future Internet Architecture.
1 Introduction
Design principles play a central role in the architecture of the Internet as driving most
engineering decisions not only at conception level but also at operational level. Many
ICT systems do not consider design principles and derive their model directly from
requirements. However, when it comes to the design of the Internet, the formulation
of design principles is a fundamental characteristic of the process that guides the
design of its protocols. On the other hand, in searching for Internet architectural
principles, we must remember that technical change is continuous in the information
and communication technology industry. Indeed, as stated in RFC 1958 [1],
"Principles that seemed inviolable a few years ago are deprecated today. Principles
that seem sacred today will be deprecated tomorrow. The principle of constant
change is perhaps the only principle of the Internet that should survive indefinitely".
In this context, it is important to provide a detailed analysis of the application of
known design principles and their potential evolution.
This paper, based on the work accomplished within the EC Future Internet
Architecture (FIArch) group [2], identifies some of the design principles that we
expect to govern the future architecture of the Internet. It may serve as a starting point
56 D. Papadimitriou et al.
and comparison basis for all research and development projects that target the so-
called Future Internet Architecture. This paper is structured as follows: Section 2
contains the definitions used in our analysis, and gives the needed background and
our understanding of the current design principles of the Internet. Section 3
summarizes the Design Principles that we expect to remain or evolve towards the
Future Internet and Section 4 gives some seeds of new design principles.
2 Definition and Background
2.1 Definitions
We define architecture the set of functions, states, and objects/information together
with their behavior, structure, composition, relationships and spatio-temporal
distribution. The specification of the associated functional, object/informational and
state models leads to an architectural model comprising a set of components (i.e.,
procedures, data structures, state machines) and the characterization of their
interactions (i.e., messages, calls, events, etc.).
Design principles refer to agreed structural and behavioral rules on how a
designer/an architect can best structure the various architectural components and
describe the fundamental and time invariant laws underlying an engineered artefact
(i.e., an object formed/produced by engineering). By “structural and behavioral
rules” we refer to the set of commonly accepted and agreed rules serving to guide,
control, or regulate a proper and acceptable structure of a system at design time and a
proper and acceptable behavior of a system at running time. Time invariance refers to
a system whose output does not depend explicitly on time (this time invariance is to
be seen as within a given set of initial conditions due to the technological change and
paradigms shifts, the economical constraints, etc.).
We use the term data to refer to any organized group of bits, e.g., packets, traffic,
information, etc. and service to refer to any action or set of actions performed by a
provider in fulfillment of a request, which occurs through the Internet (i.e., by
exploiting data communication, as defined below) with the aim of creating and/or
providing added value or benefits to the requester(s). “Resource” is any fundamental
element (i.e., physical, logical or abstract) that can be identified.
This paper refers to communication as the exchange of data (including both control
messages and data) between a physical or logical source and sink referred to as
communication end-points; when end-points sit at the same physical or logical
functional level, communication is qualified as “end-to-end”.
Security is a process of taking into account all major constraints that encompasses
robustness, confidentiality and integrity. Robustness is the degree to which a system
operates correctly in the presence of exceptional inputs or stressful environmental
conditions. Confidentiality is the property that ensures that information is accessible
only to those authorized to have access and integrity includes both “data integrity”
and “system integrity”. The term complexity refers to the architectural complexity
(i.e., proportional to the needed number of components and interactions among
components), and communication complexity (i.e., proportional to the needed number
of messages for proper operation). Finally, scalability refers to the ability of a
computational system to continue to function without making changes to the system
Design Principles for the Future Internet Architecture 57
under satisfactory and well specified bounds, (i.e., without affecting its performance),
when its input is changed in size, volume or rate.
2.2 Existing Design Principles

The Internet architecture is considered as progressively moving from the pure
network connectivity functionality to a networking ecosystem, which integrates the
network connectivity with the services combining network, computing, and storage.
Yet, beforehand, it is mandatory to establish a common understanding of the main
design principles that governs the Internet today and are mainly focused at the
connectivity functionality.
• Heterogeneity support principle [1]: heterogeneity is inevitable and must be

supported by design. This applies at many levels: devices and nodes, scheduling
algorithms and queue management mechanisms, routing protocols, levels of
multiplexing, protocol versions and implementations, underlying data link layers
(e.g., Ethernet, WiFi, etc.) and physical layers (e.g., SDH, OTN, wireless/wireline
access link), in the traffic mix and in the levels of congestion at different times and
places.
• Scalability and amplification principle: Scalability [1] states that "All designs must
scale readily to very many nodes per site and to many millions of sites". This
principle refers thus to the scale invariant that the global design should meet. The
amplification principle [3] states that “there do exist non-linearities which do not
occur at small to medium scale, but occur at large scale”. As a result complexity
can amplify small perturbations, and designers must ensure such perturbations are
extremely rare.
• Robustness and adaptability principle: the robustness principle [4] also known as
the Postel Law, warns that each protocol implementation must interoperate with
others as created by different individuals. As there may be different interpretations
of the same protocol, each one should “be liberal in what you accept, and
conservative in what you send”. This principle aims at maximizing interoperability
between protocol implementations, particularly in the face of ambiguous or
incomplete specifications. Moreover, adaptability [5] advises that “it is best to
assume that the network is filled with malevolent entities that will send in packets
designed to have the worst possible effect. This assumption will lead to suitable
protective design..."; as result, protocols would improve their robustness.
• Modularization/Layering principle: decomposes the communication functionality
into different modules with well-defined interfaces. Each of these modules
corresponds to a functional assignment which offers various behavioral and
structural advantages, such as complexity reduction, isolation, and reusability of
modules. On the other hand, modularization hinders the overall system
optimization, as each module/layer has to be optimized separately.
• Unambiguous addressing principle: following [1], the upper layer Internet protocols
must be able to identify end-points unambiguously and must be independent of the
hardware medium and hardware addressing to allow exploiting any new digital
transmission technology, and to decouple its addressing mechanisms from the
hardware. It allows the Internet to be the easy way to interconnect fundamentally
different transmission media, and to offer a single platform for a wide variety of
information, infrastructure, applications and services.
• Loose Coupling principle: Coupling is the degree to which each architectural module
relies on each one of the other modules [6]. Loose coupling defines a method for
interconnecting system components so that they depend on each other to the least
extent practicable. The extent of coupling in a system can be qualitatively measured
by noting the maximum number of element changes that can occur without adverse
effects. In today’s Internet design, “Modularity is good. If you can keep things
separate do so” [1]. The best example of loose coupling in the communication stack
is the decoupling between applicative layers and the TCP/IP protocol. The loose
coupling principle is further refined in [3] by stating that as things get larger, they
often exhibit increased interdependence between components. Much of the non-
linearity observed in large systems is largely due to coupling of horizontal and/or
vertical components. Loose coupling minimizes unwanted interaction among system
elements but can also give rise to difficulty in maintaining synchronization among
diverse components when such interaction is desired.
• Locality Principle: in computer science, this principle guiding the design of robust
replacement algorithms, compiler code generators, and thrashing-proof systems, is
useful wherever there is an advantage in reducing the apparent distance from a
process to the information or data it accesses. It has been used in virtual memory
systems, processor caches, disk controller caches, storage hierarchies, network
interfaces, etc. We distinguish the principle of temporal locality (recently accessed
data and instructions are likely to be accessed in the near future) from the spatial
locality (data and instructions close to recently accessed data and instructions are
likely to be accessed in the near future) leading to a combined principle of locality
where recently accessed data and instructions and nearby data and instructions are
likely to be accessed in the near future.
• The “end-to-end” and minimum intervention principle: End-to-end is one of the
fundamental principle on which the Internet has been structured and built, as it
guides the functional placement and the spatial distribution of functions across the
layers of the communication stack [7]. Following this principle, a function should
not be placed in the network if it can be placed at the end node (provided it can be
implemented "completely and correctly" in the end nodes except for performance
enhancement) while the core of the network should provide a general connectivity
service. The end-to-end principle has also important consequences in terms of
protocol design that should not rely on the maintenance inside the network of state
information. The application of this principle, together with the minimum
intervention (i.e., where possible, payload should be transported as received
without modification), results in a network that is transparent to the host
application communication and provides for a general, application agnostic
transport service.
• Simplicity principle: this common sense engineering principle also expressed as the
KISS (“Keep it Simple, ... Stupid”) or the “Occam’s Razor” principle, states when
facing doubts or multiple choices or ways in the design of, e.g., protocols and
intermediate systems, choose the simplest solution [1]. Adding functionality or
improving performance should not come at the detriment of increasing complexity.
• Connectionless packet switching and distributed adaptive routing: provides

unreliable, best-effort, connectionless packet delivery. The service is
“connectionless” since packets can be delivered without any prior end-to-end
connection setup phase. Forwarding decision is taken per-packet, independently at
each node: upon receiving packets, nodes lookup their routing tables to determine
the outgoing interface for that packet. The routing mechanism is called “proactive
routing” since all routing entries in the routing table must be setup before packet
delivery. Any packet can use the full link bandwidth on any link but may have to
wait in a queue if other packets are already using the link. If a datagram traverse a
hop with a full queue it is simply dropped, which corresponds to the best effort
service principle. The delivery service is thus unreliable because packets may be
lost, duplicated, delayed, or delivered out of sequence and best-effort since
delivery is not guaranteed. This switching mode also implies that it is possible to
use a stateless forwarding system at the network layer, which does not require per
connection state to ensure scalability and cost effectiveness of the communication
system and its entities.
• Network of collaborating networks - interconnection via gateways: The Internet is
often called “network of networks” since it is composed of subnets with
heterogeneous data link layer techniques and autonomous systems with
independent operation domains. Routers provide for the inter-connection of
network devices of the Internet infrastructure that is sub-divided into a collection
of autonomous systems (AS) managed by an Internet Service Provider (ISP). This
design of the routing system ensures survivability and allows for distributed
management as long as ISPs are (at least partially) collaborative.
3 Evolution of Existing Design Principles
3.1 Principles That Should Be Preserved
In this section, we detail the design principles that should be preserved and applied to
the future architecture of the Internet. Other should be adapted or augmented.
• Heterogeneity support principle: In the future, the heterogeneity is expected to be

much higher than today. Multiple types of terminals/hosts, network nodes,
protocols, and applications will co-exist. Hence, the capability to support
heterogeneity should remain (and even enforced).
• Scalability and amplification principle: the number of devices with Internet access
(e.g., computers, mobile devices), communication nodes (e.g., home, access, edge
and core routers), autonomous systems, and applications in the Future Internet is
expected to significantly increase. Moreover, the direct interconnection of the
sensor networks with the legacy Internet will exponentially increase the number of
Internet nodes. As a result, scalability is among the design principles that should
govern Future Internet, and the amplification principle would definitely remain.
• Robustness principle: the Internet is expected to increasingly handle mission and
time critical applications, related to, e.g., health, energy, and transport. As a result,
for what concerns the minimization of malfunction, uninterrupted operation and
interoperability, the robustness principle remains unchanged. Yet, as explained in

Section 3.2, this principle should be extended to cover security issues.
• Loose coupling principle: defines a necessary condition for a well-structured and
well-designed system as i) it simplifies testing and troubleshooting because problems
are easy to isolate and unlikely to spread or propagate, ii) combined with high
cohesion, it supports the general goals of high readability and maintainability, and iii)
it minimizes unwanted interaction among system components. In addition, tightly
coupled systems are likely to experience unforeseen failure states (as complex
interactions permit more complex systems to develop and make the system hard to
understand and predict) and implies that the system has less flexibility in recovering
from failure states. For these reasons, this principle shall be preserved and even
reinforced as a result of the increasing importance of the availability objective.
Nevertheless, loose coupling may also increase difficulty in maintaining
synchronization among system components when a higher degree of element
interdependence is necessary. Hence, it would be appropriate to consider that under
stress conditions, higher cohesion should be possible for proper functionality.
• Locality principle: Recent advances in computer systems engineering have pushed
cache memory to higher levels in the computer systems but the essence remains the
same: reflect the chosen methods for using the principles of spatial and temporal
locality. In this context, the locality principle should be extended to distributed
computing systems and to the higher layers space of distributed application
architectures. On the other hand, locality will play a fundamental role in self-
stabilizing distributed systems by ensure sub-linear stabilization with respect to the
number of local system components and interactions among components. As a
result, we believe that the locality principle is important and should be preserved,
while its scope should be extended to cover additional roles in distributed systems
and distributed application architectures.
3.2 Principles That Should Be Adapted (Modification of Existing Description)
In this section we highlight design principles that apply to the current Internet
architecture but should be adapted to address the design objectives of the Internet [11].
• Simplicity principle: Complex systems are generally less reliable and flexible.
Architectural complexity dictates that in order to increase the reliability it is
mandatory to minimize the number of components in a service delivery path (being
a protocol, a software, or a physical path). However, this principle has already been
challenged as complex problems sometimes require more elaborated solutions and
multidimensional problems such as the Internet architecture will be providing non-
trivial functionality in many respects. The general complexity problem can be seen
as follows: determine the placement and distribution of functionality that would
globally minimize the architectural complexity. In that respect, arbitrary lowering
complexity (over space) might result in local minimum that may be globally
detrimental. Thus, when designing the Internet, the famous quote attributed to
A.Einstein may be adopted: "Everything should be made as simple as possible, but
not simpler". Though we have to recognize that this principle is still weakly
applied, together with the conclusion of Section 3.1, scalability and simplicity
should be handled as strongly interconnected first priority design principles.
• Minimum Intervention principle: is critical to maintain and preserve data integrity
and to avoid useless intermediate information message or packet processing.
However, in some cases, it may conflict with the simplicity principle; e.g., in
sensor networks where communication gateways and actuators enable
communication between networks by offloading capabilities that would be costly
to support on sensors. As a result, we propose to relax the minimum intervention
principle as a design principle.
• Robustness principle: in order to increase robustness and system reliability, some
have advocated transforming this fundamental principle from “be liberal in what
you accept, and conservative in what you send" into "be conservative in what you
send and be even more conservative in what you accept from others". However,
adopting this approach would result in dropping a significant level of
interoperability between protocol implementations. Indeed, being liberal in what
you accept is the fundamental part that allows the Internet protocol to be extended.
With the anticipated architectural evolution of the Internet, another aspect of
interoperability will play a critical role: "how to change the engine of plane while
flying". Moreover, we shall account that the new engine can be of completely
different nature than the one it replaces. There is no universal operational principle
telling how such transition should best be performed; nevertheless it is possible to
provide the minimal conditions the new system has to support in order to facilitate
this transition. This principle however leads to relatively weak security. As stated
in [1]: “It is highly desirable that Internet carriers protect the privacy and
authenticity of all traffic, but this is not a requirement of the architecture.
Confidentiality and authentication are the responsibility of end users and must be
implemented in the protocols used by the end users”. Henceforth, we argue that the
principle should be adapted to incorporate self-protection structural principle
(coordination of the local responses to external intrusions and attacks including
traffic, data and services traceback that would enforce in turn accountability) as
well as confidentiality, integrity and authentication should be inherently offered to
information applications and services. Moreover, even if individual subsystems can
be simple, the overall system resulting from complex interactions becomes
sophisticated and elaborated. Therefore, these systems are prone to the emergence
of nonlinearity that results from the coupling between components, i.e., the
positive feedback (amplification) loops among and between subsystems and
unending oscillations from one state to another. It is possible to prevent the known
amplification loops and unstable conditions to occur but still impossible to
anticipate and proactively set the means to prevent all their possible occurrences.
In these conditions, it is fundamental to prevent propagation and that each system
keeps its own choice as last resort decision, and become "conservative to what each
system accepts and adopts".
• Modularity Principle: Current communication systems are designed as a stack of
modules structured by static and invariant binding between layers (modules) that
are specified at design time. After 30 years of evolution, communication stacks are
characterized nowadays by i) the repetition of functionality across multiple layers,
such as monitoring modules repeated over multiple layers and security components
each associated to a specific protocol sitting at a given layer (which result into
inconsistent response to attacks), which emphasizes the need to define common
functional modules; ii) the proliferation of protocol variants (as part of the same
layer) all derived from a kernel of common functions/primitives; which emphasizes
the need to define generic modules; iii) the limited or even absence of capability
for communication stacks to cope with the increasing variability and uncertainty
characterizing external events (resulting from increasing heterogeneity where
communication systems proliferate); this observation emphasizes that the
functional and even performance objectives to be met by communication systems
could vary over time (thus messages would be processed by variable sequence of
functions determined at running time); iv) the inability to operate under
increasingly variable running conditions resulting from the increasing
heterogeneity of substrate on top of which communications stacks are actually
performing. These observations lead to reformulate the modularization principle so
as to i) consider functional modules connected by realization relationships that
supply their behavioral specification, ii) distinguish between general and
specialized modules, and iii) enable dynamic and variable binding between the
different modules such that the sequence of functions performed is specified at
running time. In turn, the application of the adapted principle allows designing
systems with a larger autonomy in diagnosing internal/external stimuli but also in
their decision and execution.
3.3 Principles That Should Be Augmented (Addition to the Existing

Description)
In this section we highlight design principles that have been described and apply to
current Internet but we challenge that they should be augmented or extended.
• Polymorphism principle (as extension to the modularity principle): in computer

science/programming, polymorphism applies to data or functions. It enables to
manipulate objects of various classes, and invoke methods on an object without
knowing that object’s type. The introduction of polymorphism principle is driven
by the motivation to make use of this fact to make our architecture simpler. In
many cases, the modularity and layering principles have been the driving principles
for both communication protocols and software implementations. This principle
has led to faster deployments, but suboptimal solutions; as such these principles
have been challenged in many cases, especially in environments where functions of
each layer needs to be carried out completely before the protocol data unit is
passed to the next layer. In this context, polymorphism enables to manage and
operate first class objects belonging to different kinds of classes, while providing
the ability for a super-class to contain different objects of a subclass type at
different points in time. In turn, this allows i) for objects of different classes to
respond differently to the same function call thus results in different functionality
being executed for the same method call, and ii) for run-time (dynamic) instead of
compile-time (static) binding. Introducing polymorphism would enable the same
abstract and autonomous loosely coupled components to benefit from different
functional and/or non-functional behavior under different environments or

circumstances. The question remains open though as how to parameterize these
environmental variables and whether this could be efficiently performed through
distant exchanges (remotely).
• Unambiguous naming and addressing principle: in order to cope with the
evolution of the use of name and address spaces, the following augmentations are
considered (using [1] as starting point): i) avoid any design that requires addresses
to be hard coded or stored on non-volatile storage (when an address is an essential
requirement a discovery process is recommended); ii) A single and common
naming structure should be used. iii) Locators (LOC) and Identifiers (ID) should be
separated. In the future, it is foreseen that not only the end-points (ID) and their
attachment points (LOC) need to be unambiguous and unique within the scope in
which they appear and are used, but also the data and the services. Moreover, the
current ID/LOC approach only deals with hosts and can not provide a method to
ensure that an entity is the one claiming to be or, even worse, they disclose a fixed
identifier that can be easily traced by any other network element to know the
operations that an entity performs, thus violating its privacy.
• Extending the end-to-end principle: many experts insist that the “end-to-end”
principle is still valid, even though middle boxes and application layer gateways are
deployed at the edges of networks, as communication is divided at autonomous legs.
Another challenge concerning this principle is that IP overlay applications such as IP
multicast and mobile IP (MIP), require support from intermediate nodes (e.g., Home
Agent in MIP). It is important to notice though that some of these supports are purely
driven by arbitrary choices, (e.g., PMIP for mobility management) or delayed
migrations, (e.g., NAT instead of rolling out IPv6). Another challenge comes from
the Internet of Things, where the end-to-end communication may be significantly
modified by intermediate gateways and sensor networks sink nodes. It is also well
perceived that for many modern applications (e.g., mobile applications, distributed
searching, certain aspects of collaborative computing) maintaining state information
within the network may now be desirable for efficiency if not overall performance
effectiveness [8]. Finally, support of congestion control cannot be realized as a pure
end-to-end function: congestion is an inherent network phenomenon that in order to
be resolved efficiently require some level of cooperation between end-systems and
the shared communication infrastructure [9]. Instead of placing specific functions in
specific positions (either in end systems or routers in the network core), services and
functions must be allowed to be deployed anywhere they are needed. As a result, we
believe that motivations to "update" or augment this principle increase; however even
if this principle is challenged, it remains due to heavy consequence in terms of
scalability, survivability and robustness at large departing from this principle.
4 Seeds for New Design Principles
The Internet will evolve from a connectivity inter-network to a service ecosystem,

able to offer resources of any type (e.g., any type of network, computation, storage
and content). Realizing such Internet Architecture requires design principles that go
well beyond the networking and primitive services aspects.
In this section, we introduce seeds for completely new design principles that may
apply to the evolution of the Internet Architecture. A seed for a new design principle
refer to a concept or a notion at the inception of a well formulated design principle.
The term seed acknowledges that i) formulating principles is a complex exercise, ii)
research is still ongoing in proving their value and utility (some of our analysis and
exploitation of research results may not be mature enough) but also impact, and iii)
the proposed seeds may not be flourishing (a lot of proposal came in and very few
will materialize).
4.1 Resources Awareness
Taking into consideration that resources (associated to service components) refer to

different types (e.g., data, infrastructure resources) including resources as first order
abstraction in the Internet architecture would facilitate situation awareness. While
current service offerings are based on resource-unaware approaches, the increased
growth of both data and user-generated services poses the need for delivery schemes
(allowing media coding and rich service characterization) to overcome limitations
with regard to efficiency and quality in general. In turn, positioning as first order
abstraction processing, storage, and transmission resources (or their combination) as
well as data provides the required mechanism for the adoption of richer service-
oriented models that extend current models adapted/tailored only for higher-order
logic level(s). However, as it is also necessary for the service components network as
a whole to manage these abstractions, implies providing means to influence their
behavior.
Addressing the aforementioned challenges requires (establishing design principles
that support) the definition of suitable abstractions and mechanisms for allowing the
cooperation across all resource abstraction levels (e.g., for monitoring, negotiation).
This principle is strongly related to the “modularization principle”, and should
complement it by specifying the functional details each module exposes for
supporting crossed cooperation. Furthermore, applying this principle in combination
with the “loose coupling” principle, will allow for evaluating the effects of cross-
module awareness and cooperation, in order to avoid or minimize unwanted
interactions and non-linear effects. Another principle that needs to be considered is
the “locality” principle to which resource awareness will contribute by allowing the
development of service-delivery models enabled through self-management and cross-
module cooperation approaches.
4.2 Dependability Logic

In the current Internet there is a lack of methods and means for reliable, accountable,
and verifiable processing and handling of network and systems infrastructure with
respect to the services they host. Indeed, with the current design of the Internet:
i) Services are not cognizant of end-user expectations and needs, especially for
mission critical applications. Services are often static, lack of flexibility and they
are not negotiable. Often it is left up to the users/clients to implement their own
systems to ensure the service performs as expected
ii) Services operate on a "best-effort" basis. Moreover, services are often not
accountable towards the end-user;
iii) Services are modeled prior to their deployment in any environment and
according to the aforementioned modeling scalability rules and policies are
enforced during runtime. Nevertheless and given that infrastructures are
application-unaware, the enforced scalability rules and policies are not always
adequate to meet the application requirements in terms of efficiency,
performance, etc.; and
iv) Distributed dynamic environments ask for control policies able to deal
intelligently and autonomously with problems, emergent situations, tasks, and
other circumstances not necessarily envisaged at the design time.
The design of the Future Internet must be imbued with the principle of dependability
(reliability–accountability–verifiability feedback loop) including self-adaptation and
self-learning capability to cope and learn from changes in the operating conditions.
However, enabling such capability shall not result into monopolistic or a monolithic-
proprietary designed architecture. In that respect, this principle ought to provide
means to avoid vertical integration with proprietary components. This critical element
is part of the open research questions remaining unaddressed since so far.
4.3 Allow Exchange of Information between End-Points of Different Type
The Internet has evolved to a playground for different stakeholders such as Internet
Service Providers (ISPs), Content Distribution Network (CDN) providers, end-users,
etc. and each stakeholder tries to optimize its own utilities (or more generally
benefits), e.g., ISPs to reduce inter-domain costs, CDNs to improve content routing,
users to benefit from different choices. The so-called information asymmetry between
different stakeholders leads often the ecosystem to a suboptimal performance.
Addressing the information asymmetry problem may allow stakeholders to make
alternative decisions that would lead them collectively to a more beneficial state.
Furthermore, the emerging Design for Choice principle seed suggests that Internet
technologies should be designed so that they allow variation in outcome, rather than
imposing a particular outcome [10]. The rationale behind is that the Internet is a rather
unpredictable system and it is very difficult to assess if a particular outcome will
remain desirable in the future. The exchange of information between stakeholders
implies a flow of information from one stakeholder to another, and the “processing”
by each stakeholder; therefore the constituent capabilities of this principle include: i)
the exposure of information to a stakeholder, ii) the abstraction/aggregation of
information to be exchanged, iii) the collection of information by a stakeholder, iv)
the assessment of information by a stakeholder, and iv) the decision making.
4.4 Sustain the Resources and Brain Investment

“Coopetition” refers to the result of competing antagonistic actions due to conflicting
interests between parties implicitly cooperating in technological terms, but resulting
into negative global return - this technical term has its associated and overused
buzzword: "tussle" [10]. Instead, Internet could be designed so as to lead to a global
positive return, the so-called “all-win” situation for the society at large.
Moreover, it is important that the Internet is designed to sustain brain investment,
innovation investment and resource investment toward a global positive return. For
this purpose, it is fundamental to first recognize here the capability of the Internet to
accommodate since so far new applications communicating over a commonly shared
infrastructure (and it basically because the architecture was not designed with the idea
to privilege one class of actor against another). It is thus essential to keep the entry
barrier as low as possible and structure the design of the Internet so as to allow
various communities and people's involvement by, e.g., steer open applications
development but without impeding the genericity, evolutivity, openness, and
accessibility design objectives. Over time, the Internet shall thus cultivate the
opportunity for new players to take benefit of the infrastructure foundation without
sacrificing on its global architectural objectives and design principles. Moreover, the
Internet architecture should be able to accommodate and sustain its actors and
stakeholders’ needs in terms of fundamental capabilities, e.g., forwarding and
processing capacity.
5 Conclusion
New functionality as well as performance expectation from the Internet can be

addressed to a certain degree through incremental infrastructure investment combined
with "over-dimensioning". However, analyses have shown that increasing the
bandwidth to peta-bps on the backbone network together with system upgrades will
not suffice anymore due to new qualitative and quantitative requirements, resulting
from, e.g., highly critical services such as e-health applications, clouds of services and
clouds of sensors, new social network applications like collaborative immersive
environments, new commercial and transactional applications, new location-based
services as well as the natural expansion and growth of the Internet. Hence, a deeper
architectural evolution is required at the behavioral and structural level to sustain
these new demands that are confronted to the objective limits of the current Internet.
As design principles have played and will play a central role in the architecture of
the Internet as driving most of its engineering decisions at the conception level but
also the operational level, this document investigates their potential evolution
(adaptation and/or augmentation which arguably cover already a significant part of
their evolution). Acknowledging that new principles are emerging, this document also
explores a non-exhaustive set of new "seeds" translating current architecture research
work being realized. Altogether, the result of this investigation by the FIArch group
has lead to the identification of the design principles that will expectedly govern the
architecture of the Future Internet if corroborated by further proofs and experimental
evidences. Consequently, we believe that this work may serve as a starting point and
comparison basis for many research and development projects that target the Future
Internet Architecture. The result of these projects would in turn enable to refine the
formulation of these principles that will govern the design of the foundation of a
common architecture.
Acknowledgment. The authors acknowledge the FIArch group as this paper is a

summary of the design principles document, which is available at [2]. Part of this
work has been funded under the EC projects FP7 ICT COAST, EULER, REVERIE,
SESERV and SOFI.
References
1. Carpenter, B.: Architectural Principles of the Internet. RFC 1958 (June 1996)
2. European Commission, Future Internet Reference Architecture Group, http://ec.
europa.eu/information_society/activities/foi/docs/fiarch
designprinciples-v1.pdf
3. Bush, R., Meyer, D.: Internet Architectural Guidelines. IETF, RFC 3439 (updates RFC
1958) (December 2002)
4. RFC793, Transmission Control Protocol (1981)
5. Braden, R.: Requirements for Internet Hosts-Communication Layers. RFC1122 (October
1989)
6. Stevens, W., Myers, G., Constantine, L.: Structured Design. IBM Systems Journal 13(2),
115–139 (1974)
7. Saltzer, J.H., Reed, D.P., Clark, D.D.: End-To-End Arguments in System Design. ACM
Transactions on Computer Systems 2(4), 277–288 (1984)
8. Lyons, P.A.: The End-End Principle and the Definition of Internet. Working Group on
Internet Governance (WGIG) (November 10, 2004)
9. Papadimitriou, D., Welzl, M., Scharf, M., Briscoe, B.: Open Research Issues in Internet
Congestion Control. RFC6077 (February 2011)
10. Clark, D.D., Wroclawski, J., Sollins, K.R., Braden, R.: Tussle in Cyberspace: Defining
Tomorrow’s Internet. IEEE/ ACM Trans. Networking 13(3), 462–475 (2005)
11. Zahariadis, T., et al.: Towards a Future Internet Architecture. In: Domingue, J., et al. (eds.)
FIA 2011. LNCS, vol. 6656, pp. 7–18. Springer, Heidelberg (2011)
From Internet Architecture Research to Standards
Dimitri Papadimitriou1, Bernard Sales1, Piet Demeester2, and Theodore Zahariadis3

1
Alcatel-Lucent Bell Labs, Belgium
{dimitri.papadimitriou,bernard.sales}@alcatel-lucent.com
2
Ghent University, Belgium
[email protected]
3
Synelixis Solutions, Greece
[email protected]
Abstract. Many Internet architectural research initiatives have been undertaken

over last twenty years. None of them actually reached their intended goal: the
evolution of the Internet architecture is still driven by its protocols not by
genuine architectural evolutions. As this approach becomes the main limiting
factor of Internet growth and application deployment, this paper proposes an
alternative research path starting from the root causes (the progressive depletion
of the design principles of the Internet) and motivates the need for a common
architectural foundation. For this purpose, it proposes a practical methodology
to incubate architectural research results as part of the standardization process.
1 Introduction
The Internet model based on TCP/IP is driven since its inception by a small set of
design principles rather than derived from an architecture specification [1]. These
principles guided the structure and behavior as well as the relationships between the
protocols designed for the Internet. Nowadays, within the Internet community, some
argue that changes should be carried on once a major architectural limit is reached
(theory of change) and thus the architecture should be designed to enable such
changes. Others argue that as long as it works and it is useful for the “majority”, no
major changes should be made (theory of utility) as the objective is to keep longevity
of the design as much as possible. As a consequence of the theory of utility, the
evolution of the Internet is driven by incremental and reactive additions to its
protocols or when these protocol extensions are not possible (without changing the
fundamental properties of existing Internet Protocols) complement them by means of
overlaying protocols. Nevertheless, this approach has already shown its limits. For
instance, the design of IP multicast as an IP routing overlay led to limited Internet-
wide deployment (even if some have argued that it only enables optimizing capacity
consumption without necessarily improving end-user utility). On the other hand,
mobile IP (MIP) also designed as an IP network overlay suffers from limited
deployment too but it is undoubtedly an essential IP networking functionality to be
provided by the Internet.
In this paper, we argue that the debate between the theory of change vs. the theory
of utility is reaching its end. Indeed, the representative examples of design decisions
provided in Sections 2 aim to explain that the architecture resulting from
From Internet Architecture Research to Standards 69
independently designed components has already become the limiting factor of

Internet growth and deployment of new applications. On the other hand, the
incremental and reactive additions to protocols are becoming architecturally complex
and thus more and more time consuming; henceforth, this approach has already
reached its applicability limit too. This observation leads as explained in Section 3 to
rethink holistically the architectural foundations of the Internet itself, and thus, its
underlying research process by proposing a “third path” to architecture research. For
this purpose, we propose a method can be applied either bottom-up (results drive the
model) or top-down (model drives results). The former has also been adopted by the
EC Future Internet Reference Architecture (FIArch) Group [2]. This architecture
research initiative focuses on key architectural issues and contributes to an EC
research roadmap towards a Future Internet Architecture. From the standardization
perspective, as the Internet evolution cycle is back to research, the standardization
process has also to be reconsidered. As detailed in Section 4, the associated
challenges are i) how to best transfer architectural outcomes from research to
standard, in particular, by means of the pre-standardization process, ii) how to adapt
the standard bodies working methods to accommodate architectural research results,
and iii) how to ensure that the architectural results lead to a common baseline.
2 Architectural Model and Analysis

In this section, we provide representative examples of early decisions that drove the
design of the current Internet protocols. This non-exhaustive list of examples illustrate
perfectly well that the Internet design decisions were taken outside of any holistic
architecture albeit critical in the architectural model specification as they impact a
large portion of the Internet.
2.1 Architectural Model

Architecture is one of the key elements when engineering complex distributed
systems. Surprisingly, it has often been neglected in the context of communication
networks design, noticeably to the Internet which remains structured along relatively
weak foundations in spite of its ubiquitous deployment [3]. Many definitions of
(system) architecture have been formulated over time. In the context of this paper, we
refer to the term as “architecture” a set of functions, states, and objects/information
together with their behavior, structure, composition, relationships and spatio-temporal
distribution. More specifically, the architecture of ICT systems combines three
complementary spaces: functions: the set of procedures the system performs, their
inter-dependencies, and their relationships; objects/information: the organization of
the data the system processes (input), produces (output) by means of these functions,
including their relation and their interactions; and states: describing the system
behavior as well as the stimuli (condition, events, etc.) that change this behavior.
These spaces are modeled using formal techniques including flow block diagrams
(functions), object class combined with entity-relation diagrams (objects/information)
and finite state machines (behavior). Any “domain” of applicability ranging, e.g.,
from vending machine to avionic systems, railway signaling system, and large
ecosystems will exhibit these three complementary spaces. Hence, they also apply to
communication networks in general and to the Internet in particular.
2.2 Architectural Analysis

a) TCP connection continuity: from the beginning of the TCP/IP design, it was
decided to use the IP address as both network and host identifiers. As TCP has to
provide a reliable service, the TCP segments sent by the source to the destination are
protected by a checksum that enable to detect whether or not a segment has been
corrupted by the network. However, to ensure that a third party host can not inject
data traffic over an established TCP connection, the checksum is also performed on
the source and destination IP addresses, implying that both addresses shall remain
unchanged during the lifetime of the TCP connection. However, the IP address being
also a network identifier, the IP address of a mobile host will change together with its
attachment point raising issues at the level of the TCP connection continuity.
Resolving the latter requires a certain level of decoupling between the identifier of the
position of the mobile host in the network graph (network address) from the identifier
used for the TCP connection identification purposes.
b) IP control: IP forwarding itself is relatively simple but its associated control
components are numerous and sometimes overlapping. As a result of the incremental
addition of ad-hoc control components, their interactions are becoming more and
more complex. As of today, there is simply no systematic design of the IP control
functions which in turn causes detrimental effects such as failures, instability,
inconsistency between routing and forwarding that can lead to network black holes.
Moreover, experience shows that such practice renders the addition of control
components exponentially (architecturally) complex leading to overload of existing
components. IP routing protocols provides good examples of protocols designed with
a reduced set of kernel functions designed with limited flexibility in terms of
extension or replacement thus leading to functional overload as soon as expectations
on network functionality increases...
c) Addressing design and routing scaling: originally, host IP addresses were
assigned based on network topological location. Adoption in the mid 90's of dedicated
mechanisms to perform address aggregation (called CIDR) was felt sufficient to
handle address scaling. Today, conditions to achieve efficient address aggregation and
thus relatively small routing tables are not met anymore. This situation is exacerbated
by the current Regional Internet Registry (RIR) policy that allocates Provider-
Independent (PI) addresses that are not topologically aggregatable; thus, making
CIDR ineffective to handle address scaling. The result is that the increase of routing
table sizes worsens over time as these prefixes are allocated without taking into
account effects on the global routing system. Indeed, routing on PI address prefixes
requires additional routing entries in the Internet routing system whereas the "costs"
incurred by these additional prefixes, in terms of routing table entries and associated
processing overhead, are supported by the global routing system as a whole. Coupled
to the increase of the number of routes resulting from site multi-homing (~25% of
sites), ISP multi-homing, and inter-domain traffic-engineering, this practice
exacerbates the limitations of the Internet routing system. Nowadays, the latter must
not only scale with increasing network size and growth but also with a growing set of
constraints and functionalities. Hence, routers shall cope with increasing routing table
size even if the network itself would not be growing.
d) Border Gateway Protocol (BGP): has been designed to compute and maintain
Internet routes between administrative domains. Its route selection algorithm is
subject to Path Exploration phenomenon: BGP routers may announce as valid, routes
that are affected by a failure and that will be withdrawn shortly later during
subsequent routing updates. This phenomenon is (one of) the main reasons for the
large number of routing update messages received by inter-domain routers. In turn,
path exploration exacerbates inter-domain routing system instability and processing
overhead. Both result in delaying the convergence time of BGP routing tables upon
topology or policy change. Several mitigation mechanisms exist but practice has
shown that the reactive (problem-driven) approach at the origin of the design of these
mechanisms does not allow evaluating their potential detrimental effects on the global
routing system.
Observations: All these problems could have been avoided or at least mitigated if the
Internet was not relying on a minimalistic architectural model. Indeed, a systematic
architectural modeling of the system would have i) provided the various possible
design options from the beginning and ii) offer to the protocol designer a framework
to reason on the role of each of these components and their interactions. Without any
architecture model, the components (in particular, the protocols) tend to be designed
independently, thus, preventing any holistic approach at design time. Moreover,
independent component design does not delimit sufficient condition to achieve global
design objectives. For instance, one of the root causes of the Internet scaling resides
in the lack of modeling of the global routing system. Indeed, the main choice when
designing a routing protocol resides in the selection of the algorithm performing route
computation. However, as the routing system is not properly modeled, the impacts of
these design choices on the global routing system are almost impossible to evaluate.
In contrast, good engineering practices suggest to first model the Internet addressing
and the routing system by identifying its architectural components and their
relationships. Next, the algorithms for route computation can be designed and their
impact on the global routing system can be analyzed and evaluated by using the
architectural model. It is to be emphasized here that even if following a systematic
and holistic architectural approach does not tell the "right" routing algorithm, this
approach can certainly help delimiting what would constitute a suitable algorithm
from a functional and behavioral perspective.
What Can We Learn? The Internet architecture is implicitly defined by the
concatenation and the superposition of its protocols. In this context, architectural
components (in particular, the protocols) tend to be designed independently thus,
preventing any holistic approach at design time. Moreover, following the argument of
“utility”, the evolution of the TCP/IP model is mainly carried out by means of
incremental and reactive additions of features to existing protocols relying on a
reduced set of kernel functions. This approach has been effectively used since so far
but is now progressively reaching objective limits that range from global functional
shortcomings and/or performance degradation to maintenance1 which in turn lead to
1
Note here that the replacement and/or addition of key architectural components is impossible
without changing the properties of the architecture.
serious and costly operational problems. Hence, independent component design does
not ensure the sufficient conditions to achieve global design objectives and when
achieved lead to detrimental effects. Indeed, reasoning by protocols instead of
thinking by functions will ineluctably lead to duplicated functions, conflicting
realization of the same function and unforeseen interactions between functions
impacting the global system operation. The above examples show that the argument
of “utility” is not sufficiently compelling anymore for certain key functions such as
mobility, congestion control and routing. On the other hand, the theory of change
cannot lead to any significant improvement since there is actually no common
architectural baseline (i.e., replacement of an independent component is unlikely to
lead to a global architectural change, IPv6 being probably the best example). This
corroborates the need for conducting systematic and holistic architectural work in
order to provide a proper architectural common baseline for the Internet.
3 Architectural Foundation for the Internet
The disparity of arguments regarding the research path to follow (change vs. utility) is
resulting in maintaining the genuine Internet design foundations instead of starting
from the root causes: the progressive depletion of the foundational design principles
of the Internet. In this section, we argue that the research path to follow is not limited
anymore to the selection of the trajectory but the revision of the starting point as
determined by these root causes. We contrast the main architectural methods so as to
derive a synthetic approach that challenges these foundations.
3.1 Design Principles
Design principles refers to a set of agreed structural and behavioral rules on how an
architect/a designer can best structure the various architectural components and
describe the fundamental and time invariant laws underlying the working of an
engineered artefact. These principles are the corner stone of the Internet design
compared to architectures that rely exclusively on modeling. They play a central role
in the architecture of the Internet by driving most engineering decision at conception
time but also at the operational level. When it comes to the design of the Internet, the
formulation of design principles is a fundamental characteristic of the Internet design
process that guides the specification of the design model. On the other hand,
commonly shared design principles define necessary (but not sufficient) conditions to
ensure that objectives are met by the Internet.
Due to their importance, several initiatives have been initiated over last decade that
study the evolution of the design principles. Among others, the FIArch initiative has
undertaken a systematic analysis of the Internet design principles and their expected
evolution [4]. Analytical work on design principles documents the most common
design principles of the Internet and put them in perspective of the Internet protocols
design and their evolution. These studies aim to identify and to characterize the
different design principles that would govern the architecture of the Future Internet.
3.2 Combining Design Principles and Architectural Research

Role of Design Principles: Retrospectively, one of the most advanced architecture
for communication networks is known as the OSI Reference Model (OSI RM)
standardized in the 80's. Despite its educational value, the protocols architecture
derived from the OSI RM did not reach its expectation in terms of deployment. One
of the root causes is that the design principles regarding the OSI RM were loosely
defined; this practice resulted in lot of protocol misconceptions. Examples include the
definition of numerous options in the protocols design that renders interoperability
very challenging. This culminates with the creation of two incompatible network
layers, one based on connection-oriented and the other based on connectionless (the
so-called “CO/CL” debate). While designed at the same time than the OSI RM, the
Internet TCP/IP model and its associated protocols are nowadays used ubiquitously as
the technologies of the Internet. In contrast to the OSI RM model, the Internet model
is driven since its inception by a small set of commonly shared design principles
rather than being derived from a formal architecture. For instance, the combination of
the "end-to-end" principle with fate sharing suggested the best placement and
distribution of functionality taking into account the objective of scaling of IP inter-
networks and robustness of TCP at an acceptable cost/performance ratio.
Current Architectural Research: Inspired from [5], current approaches driving
architectural research can be subdivided into two categories:
• Driven by the theory of utility, this research assumes that the Internet shows
longevity and adaptivity thanks to its principles. Its evolution is driven at its
"edges" with the expectation to perform capabilities the network alone is unable to
provide in particular congestion control (e.g., Explicit Congestion Notification
(ECN) and its variants), and traffic-engineering (e.g., multipath-TCP) or by means
of overlays (IP multicast, mobile IP but also overlay routing and peer-to-peer fall
into this category). It is interesting to observe that independently of the investment
and research outcomes, most of these advances have had relatively limited impact
on the actual design of the Internet but also its functionality and performance.
• Driven by the theory of change, this research assumes that after several iterative
cycles of adaptation of architectural components, it becomes more effective to
redefine their foundation. Following this approach, the Internet and its design
principles are not adapted anymore to address its objectives. The architecture
resulting from reactive and incremental improvements to independently designed
protocols is already a limiting factor of the Internet growth and the deployment of
new applications (at least those that do not directly benefit from capacity addition
and/or communication system upgrades). However, in many cases, the result leads
to change/replace components as main research objective instead of resolving
architectural challenges starting from root cause analysis. A variant of this
approach assumes that the Internet can't evolve anymore because under current
conditions its design is locked by inflexible systems running processes determined
at design time to minimize the cost/performance ratio for a given set of pre-
determined functionality. Among prominent efforts falling in this category, we can
mention open-flow, and virtualization but also the more recent software-define/-
driven networks (SDN).
Third Path to Architectural Research: Following these observations, we argue that
architectural research should follow a "third path" instead of focusing on observable
consequences (theory of utility) or its premises (theory of change). This path starts by
identifying the actual root causes, i.e., the progressive depletion of the foundational
design principles of the Internet and by acknowledging the need for a common
architectural foundation relying on a revision of these principles. Indeed, without
strong motivations to adapt or to complement the current set of design principles, it is
unlikely that the current architectural model of the Internet (TCP/IP model) would
undergo significant change(s). If such evidences remain unidentified, the
accommodation of new needs either in terms of functionality or performance will
simply be realized by the well-known engineering practices residing outside the scope
of genuine architectural research work. A representative example is provided by the
evolution of the Internet communication stack that leads to reconsider the
modularization principle. This principle structures at design time the communication
stacks as a linear sequence of modules related by static and invariant bindings.
Indeed, when developed CPU and memory were scarce resources and specialization
of communication stacks for computers networks lead to a uniform design optimizing
the cost/performance ratio at design time. After 30 years of evolution, communication
stacks are characterized by: i) the repetition of functionality across multiple layers,
such as monitoring modules repeated over multiple layers (which then requires to
recombine information in order to be semantically interpretable) and security
components each associated to a specific protocol sitting at a given layer (which result
into inconsistent response to attacks), which emphasizes the need to define common
functional modules; ii) the proliferation of protocol variants (as part of the same layer)
all derived from a kernel of common functions/primitives; which emphasizes the need
to define generic modules; iii) the limited or even absence of capability for
communication stacks to cope with the increasing variability and uncertainty
characterizing external events (resulting from the increasing heterogeneity where
communication systems proliferate); this observation emphasizes that the functional
and even performance objectives to be met by communication systems could vary
over time (thus, messages would be processed by variable sequence of functions
determined at running time); and iv) the inability to operate under increasingly
variable running conditions resulting from the increasing heterogeneity of substrate
on top of which communications stacks are performing. Altogether these observations
lead to reformulate the modularization principle in order to i) connect functional
modules by realization relationships that supply their behavioral specification, ii)
distinguish between general and specialized modules (inheritance), and iii) enable
dynamic and variable bindings between the various modules such that the sequence of
functions performed is determined at running time. In turn, the newly formulated
principle provides the means to, e.g., ensure coordinated monitoring operations and
account for all security constraints (that comprises robustness, confidentiality and
integrity) consistently across all functions performed by the communicating entities.
3.3 Top-down vs. Bottom-up Method

Following Section 3.2, one can postulate that the Future Internet architecture should
rely on a revised set of design principles that governs the specification of the generic
and specialized components associated to a common architectural model [3]. Starting
from various research results (obtained by projects following the "third-path"), the
method to specify a common architecture model as defined in Section 2.1, can either
be top-down (model drives results) or bottom-up (results drive the model):
• Top-down method (see Figure 1): starts (using knowledge from research results) by
defining the global architectural level with generic and common specification
including function, information and state (step_1). Then these elements are
specialized in order to fit the needs of the domain(s) to which they apply (step_2).
By specialization we mean here the profiling of certain function and/or information
while keeping the generic properties associated to the global level. Finally these
specialized elements are translated into system level components (step_3). The
challenge here consists in specifying these components from the top so as to
produce appropriate building blocks (e.g., protocol components).
• Bottom-up method (see Figure 2): starts by exploiting research results and position
them as either global (network-level) or local (system-level). In most cases, the
corresponding elements are specialized since realized in order to reach
architectural objectives that are domain-specific. The challenge with this method
then consists in deriving from this set of common and generic components
underlying the architecture. Once identified, the result of this step is fed back in
order to align the specification of global (network-level) or local (system-level)
specific elements. Note there are no actual steps in this method that is characterized
by iterative cycles of updates between generic and specialized specification.
Fig. 1. Top-down method

Fig. 2. Bottom-up method
4 Standardization Aspects
As the evolution cycle of the Internet architecture is back to research, the

standardization process has also to reconsider its working methods i) to enable best
transfer the architectural results obtained from various research efforts to standard, in
particular, by means of the pre-standardization process, ii) to accommodate
architectural research results, and iii) to ensure that these results altogether lead to a
common architectural baseline. Considering that standardization is crucial in the wide
adoption of this baseline, this section proposes a methodology to drive its adoption.
4.1 Methodology to Guide Standardization Projects

In some cases, the standardization ecosystem related to a research project is not
ready/in place to progress its standardization objectives. In this case, a research-
focused standardization phase needs to complement the classical standardization
process to feed it with a stream of de-risked ideas that will, if successful, lead to a full
standardized solution. For this reason, this phase is generally referred to as the pre-
standardization phase. This paper develops a four-step methodology aiming at guiding
research projects to identify their standardization needs and to approach them in a
systematic way so that the necessary conditions for a successful adoption in
standardization are fulfilled:
1. Frame what needs to be standardized (interfaces, etc.) to allow the technology
proposed by the project to be interoperable and deployable at large scale. In
general, this step implies the identification of an “initial” architecture.
2. Identify the role and impact of standardization bodies on the technology segment
targeted by the project. During this step, standardization bodies are categorized as
to the role they may (or not) fulfill in the standardization ‘food chain’, i.e.,
requirements, architecture, solution/protocol, and interoperability and/or testing.
3. Evaluate the need to improve the standardization eco-system to maximize the

chance of success. This can be materialized by the creation of a new (pre-)
standardization technical committee and, in any case, requires attracting the major
stakeholders in the technology segment.
4. Identify the “structuring” dimensions (i.e., what characterizes the standardization
objectives trajectory/path) for the proposed technology/system to define a) the
criteria to shape the associated standardization target(s) of the projects, b) the
necessary conditions to meet for the technology/system to become standardisable.
The output of this step is a standardization objectives trajectory to be realized.
In this context, an (initial) architecture enables to systematically enumerate all the
interfaces and to formally analyze which of them needs to be standardized for further
transfer of the technology/system to marketable products and/or services.
4.2 What to Standardize and Where

The methodology defined in Section 4.1 is generic as it is applicable to any type of
ICT research work. When applied to the architectural research work as proposed in
this paper:
• Step 1: determine the design principles and the architectural components that
should be standardized
• Step 2: architectural work is being conducted by standardization organizations
such as 3GPP, BBF, TMF, OIF, DVB and OIPTV. Their role is to drive the
packaging of architecture solutions applicable to a given industry segment (e.g.,
wireline access, 3G mobile system, optical networks, etc.). In this context, these
architecture components are reflecting the role of the various systems involved in
the solution such as access/edge/core routers, user terminal, and eNodeB.
However, the foundational Internet architecture work we propose to conduct is
positioned as an upstream activity that will, at the end, feed these existing
architecture initiatives. The bodies where the foundational architecture work can
be standardized include IRTF/IETF, ITU-T and ETSI. More precisely, the design
principles should firstly be proposed to and evaluated together with the Internet
Architecture Board (IAB). Indeed, in terms of global reach, the most natural
place to model the architecture and its components would certainly be the IRTF.
However, IRTF (and IETF) has never considered formal and holistic architecture
work as part of its research groups charters. The ITU-T is currently working on
related thematic but not yet on components aspects. ETSI is currently hosting
several Industrial Specification Groups (ISG), some of them having a Future
Internet architecture scope. Moreover, in the context of FP7 and future EC
research programs, it would be easier to connect an ETSI ISG to the workforces
currently involved in the FIArch Group.
• Step 3: as a result of Step 2, either IRTF needs to be convinced and willing to
step into holistic architectural work, or the current ITU-T work program needs to
be reinforced/refocused or new work item needs to be launched within ETSI.
• Step 4: it is unlikely that a standardization body will accept to incorporate the

architectural aspects of the research output ‘as-is’ in its standardization work
program (cf. structuring dimensions of Step 4), mainly because this architectural
work needs further validation. As a result, it is proposed to start the work in the
pre-standardization mode following either the top-down or the bottom-up method
defined in Section 3.3.
Fig. 3. Application of the bottom-up method
The objectives of the pre-standardization phase will be i) to validate a possible

reference architecture model for the future Internet, ii) to give guidelines on
alternative models (if applicable), and iii) to explore possible directions for further
standardization.
We consider hereafter setting up a potential pre-standardization group (PSG)
dedicated to the Future Internet architecture and implementing the iterative bottom-up
method as described in Section 3.3. In the context, such a PSG (that could be initiated
at, e.g., ETSI or IRTF) will follow the working method proposed in this paper:
• The design objectives and principles provided by the EC FIArch group will be
used as input and will drive the specification of the model and components part
of the architecture. However, the industry and academia at large will have also a
way to influence them in order to ensure their broad acceptance.
• Research projects and academia will contribute to the architecture work in their
domain of expertise; these inputs will be used to build an acceptable architecture
per key domain, e.g., sensor networks/Internet of Things (IoT), networked-media
and data-centers/cloud.
• Common building blocks shared between these domain-specific architectures will
be identified; next, the domain-specific architectural components will be
aggregated to create the generic building blocks and the relationship between
these blocks will be determined.
4.3 Steps toward Architectural (Pre-)Standardization

Standardization work is generally driven by the contributing efforts of its participants.
Pre-standardization work on the Future Internet architecture will not depart from this
mode of operation. Note that concerning the generic architecture work, the effort will
be conducted within the architecture pre-standardization group (PSG). In any case,
two distinct scenarios could be envisaged either in the global and/or the domain-
specific architecture context:
• First scenario: when a core contribution is submitted to the PSG, it will naturally
serve as the fundamental basis for the further discussions. This will happen, for
instance, when a team agrees on a given architecture outside the PSG and submits
it as input to the pre-standardization process. In this case, contributions from
other participants will be in most cases limited to improvements with respect to
the core proposal, consensus will be easier to reach and the process will converge
quickly to a consolidated architecture reflecting the view of the PSG participants.
• Second scenario: two (or more) competing proposals are brought to the PSG. In
this case, reaching consensus within the group becomes more challenging. The
way to resolve potential conflicts at the architectural level includes i) the
identification of common components (from the competing proposals) fulfilling a
completely or partially similar role with respect to the architecture and organize
the model accordingly, and ii) the organization of the components that are
complementary around a kernel of common components. It is also possible that
competing models are actually complementary, leading to an architecture where
both models are loosely “interconnected” with a few number of data and/or
function relationships depending on the model being specified.
The architecture PSG will be chartered for a limited lifetime. When the global/generic
and domain (both local and global) architectural models produced by the PSG are
validated and considered as mature enough, the PSG work will have to be transferred
to the normal standardization process. The global/generic architecture work can be
standardized by the body hosting the PSG. However, concerning the domain-specific
(both local and global) architecture work, the standardization body hosting the PSG
may not cover all technical domains. As a result, for some domains, the work needs to
be reassigned to another standardization body, thus inducing the creation, within the
targeted standardization bodies, of one or more architecture working groups/work
items to work on a first standardized version of the domain specific architecture.
Then, any proposal for new technology, solutions, or protocols will have to be
positioned with respect to the existing architecture, including a clear analysis on the
impacts of the new proposal on the architecture. If necessary, changes to the existing
architecture should be identified and clearly motivated. It should be noted that the
architecture will also drive further specification of protocols in order to realize the
implementation of its identified interfaces. This new architecture-centric approach is
expected to realize in the standardization context the “Third Path to Architectural
Research” proposed in this paper (cf. Section 3.2).
5 Conclusion
In this paper, we argue that the debate between architectural research driven by the
application of the theory of utility and the theory of change is over. Indeed, neither of
these approaches can fundamentally address the limits of the Internet architecture.
Instead, we propose that architectural research should follow a "third path" starting by
identifying the actual root causes (by adapting the foundational design principles of
the Internet such as the modularization principle) and by acknowledging the need for
a holistic architecture (instead of (re-)designing protocols independently and
expecting that their combination would lead to a consistent architecture at running
time). The proposed path will in turn also partially impact how the necessary
standardization work is to be organized and conducted. Indeed, the design principles
and the modeling part of the architecture need to be standardized to ensure its
adoption at the international level. Following this path, the chartering of the new work
item to define, e.g., new protocol, will need to be not only “problem-driven” but also
“architecture driven”. It is also anticipated that, resulting from the current wave of
Future Internet research projects, the pre-standardization work will become more and
more relevant with a mix of architecture- and technology-driven work items. As such,
this is an opportunity since this nascent pre-standardization ecosystem can be seen as
a laboratory to learn how to introduce an “architecture-driven” dimension in the
Internet standardization working method.
References
1. Carpenter, B.: Architectural Principles of the Internet. IETF, RFC 1958 (June 1996)
2. European Commission, Future Internet Reference Architecture Group (FIArch), http://
ec.europa.eu/information_society/activities/foi/research/
fiarch/
3. Papadimitriou, D., Sales, B.: A path towards strong architectural foundation for the
internet design. In: 2nd Future Network Technologies Workshop, ETSI, Sophia Antipolis,
France, September 26-27 (2011)
4. FIArch Group Report v1.0. Future Internet Architecture-Design Principles (January 2012)
(in press)
5. Clark, D.: Designs that last. In: FP7 EIFFEL Think Tank, Athens, Greece, October 6-7
(2009)
6. Zahariadis, T., et al.: Towards a Future Internet Architecture. In: Domingue, J., et al. (eds.)
FIA 2011. LNCS, vol. 6656, pp. 7–18. Springer, Heidelberg (2011)
An Integrated Development and Runtime
Environment for the Future Internet
Amira Ben Hamida1 , Fabio Kon5 , Gustavo Ansaldi Oliva5 ,

Carlos Eduardo Moreira Dos Santos5 , Jean-Pierre Lorré1, Marco Autili4 ,
Guglielmo De Angelis6 , Apostolos Zarras3, Nikolaos Georgantas2,
Valérie Issarny2 , and Antonia Bertolino6
1
Petals Link, France
2
Inria, France
3
University of Ioanina, Greece
4
Università degli Studi dell’Aquila, Italy
5
University of São Paulo, Brazil
6
Institute of the National Research Council of Italy
Abstract. The Future Internet environments raise challenging issues

for the Service-Oriented Architectures. Due to the scalability and hetero-
geneity issues new approaches are thought in order to leverage the SOA
to support a wider range of services and users. The CHOReOS project
is part of the European Community Initiative to sketch technological
solutions for the future ultra large systems. In particular, CHOReOS
explores the choreography of services paradigm. Within this project, a
conceptual architecture combining both the development and runtime
environments is realized. This chapter introduces the CHOReOS Inte-
grated Development and Runtime Environment, aka IDRE.
Keywords: SOA, Service, Choreography, MDA, Cloud & Grid, IDRE,

Governance, Middleware, Discovery, Access, TDD, V&V.
1 Context
Raising the Future Internet Challenges. The Future Internet (FI) context draws
a global environment populated with a plethora of services. Such services are
related to two - commonly identified by many FI initiatives - key FI dimensions,
the Internet of (traditional) Services and the Internet of Things. The latter di-
mension is expected to considerably change the way we perceive the Internet
today, by incorporating in it vast populations of physical objects or, from an-
other viewpoint, sensors and actuators linking to the physical world. We take
this SOA view of the FI one step forward by advocating choreographies of ser-
vices i.e., compositions of peer interacting services as the primary architectural
solution for leveraging and sustaining the richness and complexity of the FI. In
this context, three key challenges, namely, scalability, heterogeneity, and aware-
ness are raised. As already pointed out, the large scale of today’s Internet be-
comes ultra large scale (ULS) in the FI, in terms of numbers of devices, services,

82 A. Ben Hamida et al.
things, users, requirements, and their infinite combinations within choreogra-

phies. Then, extreme heterogeneity is unavoidable in terms of the previous, and,
additionally, in terms of interaction protocols at different levels, data, semantics,
and related technologies. Third, awareness has to do with taking into account
user requirements as well as context in all its dimensions, physical, system, and
user context, as well as its volatility in the open, dynamic and mobile FI. The
CHOReOS project is part of the European Community Initiative to sketch tech-
nological solutions for the future ultra large systems. In particular, CHOReOS
explores the choreography of services paradigm.
Addressing the Future Internet Challenges. In this paper, we provide a com-

prehensive solution to the above particularly challenging issues. We realize the
CHOReOS Integrated Development and Runtime Environment, aka IDRE. We
exploit sophisticated research domains from the Service-Oriented Architecture
(SOA) realm, including Service Discovery, Access and Composition as well as
SOA Governance, together with the Model Driven Engineering (MDE) paradigm
and the Cloud & Grid paradigms [9]. Building on MDE principles, the CHOReOS
development process enables going from very high-level user requirements for ser-
vice choreographies down to highly heterogeneous realizations of the final chore-
ographies, where incompatibilities of the participating services are compensated
for. It is worth noting that in the CHOReOS terminology (traditional) Ser-
vices and the Internet of Things, become, respectively, the Internet of Business
Services (IoBS) and the Internet of Thing-based Services (IoTS). To deal with
environments where IoBS and IoTS coexist in a transparent way, CHOReOS
IDRE relies on the integration, interoperability and large scale distribution ca-
pabilities provided by the Enterprise Service Bus middleware paradigm, which
we extend and enhance to cope with the very heterogeneous deployment and
interaction semantics and platforms of both types of services. Additionally, we
develop sophisticated service discovery mechanisms in order to offer registration,
classification, query and retrieval mechanisms adapted to the ULS populations
of Business Services and Thing-based Services. Scalability issues are also consid-
ered at the levels of service access and provisioning, choreography deployment
and need for computation, as well as management of vast populations of services
and their data, where we exploit Cloud and Grid capabilities for offering a pow-
erful and elastic platform of resources. Finally, we rely on the fundamentals of
the Governance and Verification & Validation (V&V) domains for ensuring the
quality of services and choreographies at both design and run time. Both func-
tional and non-functional properties of services and choreographies are assessed,
augmenting in this way our awareness of the composed choreographies. In this
chapter, we introduce the IDRE conceptual view, detailing its subsystems and
their respective functionalities. The remainder of the chapter is as follows. We
provide an overview of the CHOReOS IDRE in Section 2. Section 2 is dedicated
to the choreography synthesis. In Section 4, we detail the CHOReOS middle-
ware. Section 5 introduces the Governance and V&V framework. Finally, we
conclude in Section 6.
An Integrated Development and Runtime Environment for the FI 83
2 CHOReOS IDRE Overview

The CHOReOS IDRE relies on a modular SOA where a number of coarse-grained
subsystems are integrated to support the overall development, from design to
implementation, together with deployment and execution, of services choreogra-
phies in the FI. CHOReOS embeds the following subsystems: the CHOReOS
Development Environment, the eXecutable Service Composition (XSC), the eX-
tensible Service Discovery (XSD), the eXtensible Service Access (XSA), Cloud
and Grid Middleware, Governance and V&V Framework and finally the Moni-
toring (See Figure 1).
Fig. 1. CHOReOS IDRE Overview
3 CHOReOS Development Environment

ULS choreographies of services need to be created and analyzed with the aim of
coping with the FI environments. For this purpose, the CHOReOS project pro-
vides a dedicated development environment (See Figure 2). Therefore, a model-
driven development process is realized. First, thanks to dedicated Requirements
Specification Tools the user requirements specification is captured. The final
output of the requirements specification activity is a choreography specification
(in the BPMN2.0 language), which serves as input to the next phases of the
overall process. Second, the Synthesis Processor operates an automated synthe-
sis of specific software entities, namely Coordination Delegates, that coordinate
the collaboration among the services so as to enact the choreography in a fully
distributed way. These are executed on top of the CHOReOS Middleware (See
Section 4). Third, the development process ends with the scalability analysis
performed by the Choreography Analyzer.
Fig. 2. CHOReOS Development Process
3.1 Requirements Specification Tools
The Requirements Specification Tools are mainly responsible for enabling do-
main experts to specify functional and quality requirements on services and
service-based applications, and in turn, to enable the domain expert to produce
a first draft choreography specification. First, the Specification Expressing Tool
and DataBase provide the domain expert with service consumer requirements
and associated attributes. The service consumer specifies requirements using a
structured approach facilitated by mobile tools – such as the iPhone app (ap-
plication). There can be many service consumers with many user needs. The
expressed requirements are recorded in a DataBase along with attributes for
quality, priority and situation. Associated with the requirements is a quality
model, which relates the user requirements on service-based applications to QoS
on services aggregated in these applications. Second, the Requirements Manage-
ment and Analysis Tool provides the domain expert with requirements manage-
ment and analysis functions. These functions are provided to help the domain
expert to pull out individual requirements in order to form a set of require-
ments for choreography. Third, the Requirements Engine executes a matching
and grouping algorithm to cluster the service consumer and domain expert ex-
pressed requirements. A ‘calculate similarity’ algorithm, enables the requirement
comparison for similarity using natural language processing techniques. The out-
put from this component is grouped requirements for choreographies. Finally,
the Matching Tool and User Task Model Database are responsible for matching
the requirements on the choreography specification to user task models using a
matching tool. A set of CTT (Concur Task Trees) task models, describing struc-
tured activities that are often executed during the interaction with a system are
defined and stored in a database. Finally, the prioritized quality-based require-
ments and user task models are then associated with choreography strategies,
which are expressed in the form of patterns by the choreography designer. The
final output of this process is a first draft choreography specification and a set
of associated requirements to inform the discovery of abstract services.
3.2 Synthesis Processing
Advancing the foundational background on software coordination via automated

coordinator synthesis [3,8,11], the Synthesis Processor subsystem is mainly re-
sponsible of synthesizing the coordination delegates that are in charge of suitably
coordinating, in a distributed way, the services participating to the choreogra-
phy. The approach starts from the BPMN2 choreography model and from the
set of discovered services. The first input comes from the refinement of the CTT
models and choreography patterns (and hence, the first draft choreography spec-
ification discussed in Section 3.1). The latter comes from the exploitation of the
service base management mechanisms described in Section 4.1. Thus, the synthe-
sis process assumes that the services into the registry/base have been discovered
so that they satisfy the local (to the service) functional and non-functional re-
quirements that have been specified for the choreography and, hence, can be
considered as potential candidates to participate in the global choreography
process. Finally, the choreography synthesis produces the coordination delegates
that will be then managed by the service composition engine for choreography
realization purposes presented in Section 4.2, hence accessing the participant
services through the service access subsystem presented in Section 4.3.
3.3 Choreography Analyzer

Given the ultra large scale of FI choreographies, automated analysis mechanisms
become necessary to support choreography evolvability. The Choreography Anal-
ysis component is mainly responsible for analyzing either a serialized BPMN2
choreography specification or the set of coordination delegates issued by the
synthesis process (Section 3.2). Two kinds of analysis are currently supported
and implemented in the form of subcomponents, namely choreography Scalabil-
ity prediction and choreography Stability and Interdependencies Analysis. In the
following, we describe each of these subcomponents. The Scalability Prediction
relies on two mechanisms: the QoS Prediction and the Scalability Analysis. The
QoS Prediction aims at estimating the behavior of the choreography (written in
BPMN2.0) regarding QoS parameters such as service response time, capacity,
reliability, availability of a composition, etc. The prediction takes into account
the choreography execution context (the number of user requests, the number of
concurrent choreographies, the available resources), but captures it in a single
state. In turn, the Scalability Analysis considers various possible states of the
choreography execution. It uses for this issue the QoS Prediction mechanism for
single state prediction and a mathematical model describing the dynamics of
changes in the choreography execution. The Stability and Interdependency An-
alyzer is primarily responsible for performing change impact analysis based on
the existing dependencies between choreography participants. In addition, the
component also applies the analysis to a set of concrete services and coordina-
tion delegates that realizes the choreography. The analyzer component relies on
model-to-model (M2M) transformations to obtain the dependency graph from
either a choreography BPMN2.0 specification or a set of coordination delegates.
Finally, the analyzer relies on graph analysis techniques to calculate a variety of
dependency-centric measures, including graph centralities [10] and stability.
4 CHOReOS Service-Oriented Middleware

The CHOReOS middleware targets two different but interrelated domains of
services: Business services and Thing-based services. Based on this inherent
characteristic, the high-level architecture of the CHOReOS middleware com-

prises corresponding domain-specific mechanisms that support the discovery
of services, the access to services, and the execution of service compositions.
The specificities of the functionality offered by the domain-specific mechanisms
are hidden by corresponding unified “eXtensible” middleware mechanisms that
unify the access to the domain-specific middleware mechanisms. In addition,
computationally- and storage-intensive tasks of both the middleware and the
choreographies are supported by the CHOReOS Cloud and Grid services. In the
following, we describe the CHOReOS middleware (See Figure 3).
Fig. 3. CHOReOS Middleware Architecture
4.1 eXtensible Service Discovery

The CHOReOS IDRE provides a multi-protocol service discovery service. Actu-
ally, it relies on an Abstraction-oriented Service Base Management (AoSBM) [2],
[1] that stores and classifies in a suitable way an important amount of services
data. This base is populated by an extensible plugin-based mechanism. The lat-
ter is responsible for extending the service discovery to both business services
and FI things, by plugging domain-specific discovery protocols like, e.g., the
Governance Registry for Business services and the Things Discovery Protocol
(TDP). The registry is populated either in a passive or active way thanks to
the Plugins. Moreover, the XSD relies on Semantic Knowledge Base ontologies
to enable the devices (Things) discovery. Regardless of their heterogeneity, the
CHOReOS XSD provides a unique abstraction referencing services, which em-
powers the ability of realizing choreographies of services.
4.2 eXecutable Service Composition

Service choreographies in CHOReOS are supported by different execution plat-
forms so as to cope with the diversity of service technologies found in a ULS
environment. The enactment of a choreography is taken into account by the
XSC mechanism. The latter takes as input the choreography synthesized previ-
ously by the Choreography Development Process and discussed in Section 2. A
set of Coordination Delegates specifies the choreography and are then dedicated
to the right XSC. Specifically, BPEL-based XSC enables the implementation
of coordination delegates using BPEL, while SCA-based XSC supports the im-
plementation of coordination delegates using SCA. In a complementary way,
the Thing-based Composition & Estimation component deals with the composi-
tion of Thing-based services to handle requests for interacting with the physical
world. While enacting a choreography, some services may not respect the ini-
tially contracted agreements and choreography reconfigurations need then to be
operated. For that end, the CHOReOS XSC relies also on a reconfiguration and
substitution mechanism.
4.3 eXtensible Service Access

In ULS environments, millions of services either Business or Things oriented are
deployed in a distributed manner allover the ambient context. Besides the fact
that they are coming from heterogeneous sources and are dedicated to different
aims, they are also implemented using distinct standards and technologies. Con-
sequently, in order to make these services collaborate within choreographies, it
is essential to provide a middleware technology that unifies their access. Within
CHOReOS we exploit the Enterprise Service Bus paradigm, which provides a
glue technology supporting connectivity and communication techniques. Further
enhancements are realized with the aim of making the CHOReOS XSA support
both Business and Things services. Indeed, the XSA is based on an enhanced ser-
vice bus paradigm to overcome the heterogeneity of the FI. This paradigm is rep-
resented by the eXtensible Service Bus (XSB). The latter enables multi-protocol
access to both Business Services of the IoBS domain and Thing-based Services of
the IoTS domain, as well as cross-domain access. In particular, it enables inter-
operability among heterogeneous interaction paradigms of both domains, while
conserving as much as possible their semantics. The XSB is an abstract bus that
prescribes only the high-level semantics of the common bus protocol. This se-
mantics follows a Generic Application-GA- abstraction paradigm. Moreover, the
XSB relies on the Distributed Service Bus (DSB) [4] that provides support for
accessing business services. We rely on the Petals DSB to ensure this function-
ality. Additionally to the native bus capabilities the DSB supports distribution
and offers the core runtime middleware. The DSB is leveraged to the FI features
through the adaptation to the cloud middleware. It benefits from the provided
hardware resource infrastructure, in fact. Further enhancements intend to take
advantage from the cloud elasticity discussed in Section 4.4. Furthermore, in or-
der to target IoTS domain, we provide the Light Service Bus (LSB), which is a
lightweight concrete bus realization of XSB and its GA semantics, dedicated to
IoTS, hence, accounting for its dynamics and resource constraints while enabling
access to heterogeneous Things. In particular, the GA semantics is conveyed on
top of a substrate protocol (DPWS) that is suitable for the IoTS domain.
4.4 Cloud and Grid Middleware

The Cloud and Grid Middleware services provide basic services that support
computational- and storage- intensive tasks performed either by the CHOReOS
middleware services, or by the choreographies that are built on top of the

CHOReOS middleware. The Cloud service can allocate and deallocate resources
dynamically according to service demand. Tasks such as encoding large amounts
of video in a citizen journalism application can take advantage of the Grid
service. The allocation of Cloud machines for execution of choreographies is
performed by the CHOReOS middleware in a way that is transparent to chore-
ography users, designers, and developers. The CHOReOS middleware uses the
ServiceDeployer component to allocate new nodes from the NodePoolManager
and then deploy and run new services on them. In these nodes, CHOReOS will
execute major choreography components (e.g., proxies, adapters, coordination
delegates) for service access at runtime. To this end, the EnactmentEngine will
use the NodePoolManager and the ServiceDeployer to set up the choreography
environment, allocating the required nodes, deploying the required software and
enabling the execution of the choreography. To achieve scalability and portabil-
ity, the NodePoolManager is able to allocate new nodes in multiple underlying
execution platforms. A CHOReOS node may be part of a Cloud Infrastructure
as a Service (IaaS) platform; these can be provided by a public Cloud such as
Amazon EC2 or Rackspace, or a private Cloud, for example, executing the Open-
Nebula or OpenStack open source Cloud software. The CHOReOS monitoring
service will provide data to runtime QoS and V&V enforcers. If a QoS violation
is detected, for example, the Cloud service can be used to allocate new nodes in
an attempt to improve QoS.
5 Governance and V&V Framework
ULS choreographies bring into play a very large number of services, users and
resources employing the system for different purposes. Therefore, methodologies
and approaches that will permit the smooth integration of independently devel-
oped pieces of software need to be implemented. In IT Systems, the Governance
approach enables supervising such large systems. Indeed, a set of processes, rules,
policies, mechanisms of control, enforcement policies, and best practices are put
in place throughout the life-cycle of services and choreographies, in order to
ensure the successful achievement of the SOA implementation. Activities such
as policy definition, auditing & monitoring, and finally evaluation & validation
are recommended. Within CHOReOS, we implement a Governance and V&V
Framework (See figure 4) that underly the services, and choreographies lifecy-
cle. Precisely, the Service Level Agreement-SLA and lifecycle management deals
with the lifecycle of relevant resources such as services, service level agreements,
and choreographies. Further, the V&V Components perform the testing of ser-
vices before their involvement in choreographies. Online testing of services and
choreographies at runtime is also operated. Finally, the Test Driven Development
Framework (TDD) operates a series of complementary tests.
Fig. 4. CHOReOS Governance and V&V Framework
5.1 SLA and Lifecycle Management

The SLA and Lifecycle management activities [12] are responsible for offering
the capabilities that ease the management of the resources, these can be services,
choreographies, policies, and service level agreements life-cycles. Meanwhile, it
also helps managing the roles and responsibilities of the users of the Gover-
nance Framework, by assigning credentials. Design time policies might define
which, when, and where to use standards and insure compliance between them.
The design time policies may also consider the fact of setting out corporate
namespaces, common coding conventions, etc. When a choreography is enacted,
in order to ensure its good behavior, it is ultimately necessary to enforce and
manage the non functional contracts of involved services, according to defined
policies. Within CHOReOS, a distributed monitoring system is envisaged in or-
der to assess the ULS choreography properties. The SLA governance is realized
by the Business Service Monitoring discussed in Section 5.4.
5.2 Verification and Validation Support

The Governance framework provided by the CHOReOS project implements a
comprehensive strategy for managing both choreographies and services. The
project put a special emphasis on governance aspects related to choreography-
oriented V&V activities by defining policies, and rules governing (e.g., enabling,
regulating, etc.) them [6]. The idea of V&V governance was originally proposed
in [7] to support an on-line testing [5] session when a service asks for registration
within a registry. In this vision, only services passing the testing phase are logged
in the registry. As a result, the registry is expected to include only “high-quality”
services that passed the validation steps foreseen by a more general governance
framework. In addition to the registration of a new service, the on-line validation
process could be also extended to other events, like the release of a new service
version. Note that when entering a new service registration in a registry, the
service provider is naturally wishful to promote the service and therefore can be
explicitly willing to submit it to on-line testing. On the other hand, the notifi-
cation of a service upgrade could be notified only sporadically. The governance
mechanisms oriented to V&V activities could mitigate this aspect by means of

specific policies and obligations that the service providers should abide by, when
binding their services to a choreography. During the life-cycle of a choreography,
a service that was originally registered to play a given role in such choreography
could be modified or become deprecated. In addition, it is also possible that a
single service may play one or more roles defined by a choreography. Finally, the
same service may be involved in several choreographies with differnt roles, as well.
In all these scenarios, the V&V governance rules that the CHOReOS project is
proposing aim at prescribing that any modification (i.e. activation, modification,
cancellation) to either a registered service, or to a role defined by a choreography
should activate a new appropriate online testing session. In this sense, the Gov-
ernance Registry is an important component for SOA Governance. Indeed, as
described above, the CHOReOS Governance Framework enhances the canonical
functionalities provided by a Service Registry with a set features supporting on-
line testing techniques. Specifically, each feature is implemented and managed
by proxing the Service Registry with a set of dedicated handlers. Such handlers
are conceived as mechanisms permitting to modify the registration procedure of
a service with additional functionalities. In particular testing handlers activate
testing sessions on services for which a registration request, or a modification of
the associated entry, is received.
5.3 Test-Driven Development Framework

The main goal of Rehearsal, the CHOReOS testing framework, is to support
Test-Driven Development (TDD) of web service choreographies. Using the frame-
work, a choreography developer can perform multiple levels of tests to guide the
choreography development. TDD is performed in a testing, or offline, environ-
ment where some of the concrete services may not be available. To achieve that,
Rehearsal provides mechanisms for emulating real services or a part of the chore-
ography by using mocks, which is a well-known TDD practice. In addition, the
framework provides mechanisms for applying unit, integration, and scalability
testing. At development-time, services may be created or adapted to implement
the choreography roles properly. Unit testing aims at validating the correct be-
havior of atomic services. Integration testing aims at validating the messages
exchanged by the services when they are composed to implement a role. Finally,
compliance tests may also be applied to verify whether a service or a composition
of services plays the role correctly. Rehearsal also supports the scalability testing
of choreographies. Using this feature, the developer can assess the choreographies
in different scales. Through this assessment, which is performed offline, the de-
veloper can estimate the needed infrastructure aspects (e.g., instances of virtual
machines allocated to a service) to assure a performance metric (e.g., response
time) in the online environment. As it is a framework, Rehearsal usage, itself,
does not imply TDD application. It must be composed with other classes to
create a concrete and executable application. For so, a methodology is proposed
to guide developers in the application of TDD in choreography development
using Rehearsal. This methodology is divided in four phases: (i) Creation and
adaption of atomic services; (ii) Integration of services to compose choreography

roles; (iii) Integration of roles to compose the choreography; (iv) Acceptance
and scalability testing. The framework provides a tool to support each of these
phases. All tests written using TDD serve both as an executable specification of
the choreography behavior and as a means for V&V at design time. Later, at
runtime, the same tests may be used with the online system to verify the proper
behavior of the choreography in the production environment.
5.4 CHOReOS Monitoring

Relevant data such as the functional as well as the non functional attributes of
services are useful for the system supervision. The data from hardware resources
helps the middleware to engage in reactive measurements to correct problems as
they occur. However, monitoring ULS choreographies and systems raises chal-
lenging issues such as dealing with the scalability, the distribution and the het-
erogeneity. The CHOReOS IDRE addresses these requirements by relying on a
distributed and event-based infrastructure for monitoring both Business Services
and hardware resources. Finally, a Complex Event Processor-CEP-, ensures the
respect of the dictated policies. Once services are deployed or exposed on the
DSB, the CHOReOS Monitoring performs runtime assessment of Service Level
Agreements and control of the communications taking place within a choreogra-
phy, thanks to the Business Service Monitoring, which gathers data from the Dis-
tributed Service Bus. Communication Monitoring is achieved by subscribing to
events triggered by the running services. While, QoS Runtime Assessment relies
on the implementation of the WSDM standard. Then, the Resource Monitoring
is a ganglia-like monitoring system that interacts with the Cloud and Grid Mid-
dleware. First, it actively supplies notifications to interested subsystems about
relevant events, such as overloaded systems, out-of-memory conditions, or hard-
ware failures. Second, it maintains an overview of the current and recent status
of system resources to be able to respond to queries about them. Queries are
useful to support creation or destruction of virtual machine instances according
to load, services allocation, or services migration. We address the scalability and
distribution issues by considering for each each CHOReOS node a local compo-
nent collecting data (primarily memory, disk, and CPU usage). Then, data is
aggregated between distributed nodes in a hierarchical manner.
6 Conclusion
The FI world challenges the SOA by raising scalability, distribution and het-
erogeneity issues. The CHOReOS project addresses these issues by providing
responses at several levels. The CHOReOS Integrated and Runtime Environ-
ment gathers top-level technological SOA approaches including Model-driven
Architectures, SOA Discovery, SOA Composition and SOA Governance. In this
chapter, we have presented the CHOReOS platform as well as its main compo-
nents. Ongoing works concern the realization of the IDRE in ULS choreography
use cases.
Acknowledgments. This work is partially funded by the EU 7th Framework

Program, contract CHOReOS FP7-ICT-2009-5 - Objective 1.2 (grant 257178).
The work presented in this paper only reflects the authors views. The Commision
is not liable for any use that may be made of the information contained therein.
credited.
References
1. Athanasopoulos, D., Zarras, A., Issarny, V.: Towards the maintenance of service
oriented software. In: Proc. of the 3rd CSMR Workshop on Software Quality and
Maintenance, SQM (2009)
2. Athanasopoulos, D., Zarras, A., Vassiliadis, P., Issarny, V.: Mining service abstrac-
tions - nier. In: Proc. of the 33rd International Conference on Software Engineering
(ICSE), pp. 944–947 (2011)
3. Autili, M., Mostarda, L., Navarra, A., Tivoli, M.: Synthesis of decentralized and
concurrent adaptors for correctly assembling distributed component-based systems.
Journal of Systems and Software (2008)
4. Baude, F., Filali, I., Huet, F., Legrand, V., Mathias, E., Merle, P., Ruz, C., Krum-
menacher, R., Simperl, E., Hamerling, C., Lorré, J.: Esb federation for large-scale
soa. In: Proc. of the ACM Symposium on Applied Computing, SAC 2010, pp.
2459–2466 (2010)
5. Bertolino, A., De Angelis, G., Kellomäki, S., Polini, A.: Enhancing service federa-
tion trustworthiness through online testing. IEEE Computer 45(1), 66–72 (2012)
6. Bertolino, A., De Angelis, G., Polini, A.: Validation and verification policies for
governance of service choreographies. In: Proc. of the 8th International Conference
on Web Information Systems and Technologies, WEBIST (to appear, April 2012)
7. Bertolino, A., Polini, A.: Soa test governance: Enabling service integration testing
across organization and technology borders. In: Proc. of Software Testing, Verifi-
cation and Validation Workshops (ICSTW), pp. 277–286 (April 2009)
8. Calvanese, D., De Giacomo, G., Lenzerini, M., Mecella, M., Patrizi, F.: Automatic
service composition and synthesis: the roman model. IEEE Data Eng. Bull. 31(3),
18–22 (2008)
9. Issarny, V., Georgantas, N., Hachem, S., Zarras, A., Vassiliadis, P., Autili, M.,
Gerosa, M., Ben Hamida, A.: Service-Oriented Middleware for the Future Inter-
net: State of the Art and Research Directions. Journal of Internet Services and
Applications 2(1), 23–45 (2011)
10. Newman, M.: Networks: An Introduction, 1st edn. Oxford University Press (2010)
11. Tivoli, M., Inverardi, P.: Failure-free coordinators synthesis for component-based
architectures. Sci. Comput. Program. 71, 181–212 (2008)
12. Zribi, S., Bénaben, F., Ben Hamida, A.: Towards a service and choreography gov-
ernance framework. In: Proc. of the I-ESA Conference, Valencia Spain. Springer,
Heidelberg (to be published, 2012)
Visual Analytics: Towards Intelligent Interactive
Internet and Security Solutions
James Davey1 , Florian Mansmann2 , Jörn Kohlhammer1 , and Daniel Keim2

1
Fraunhofer IGD, Germany
2
Universität Konstanz, Germany
Abstract. In the Future Internet, Big Data can not only be found in
the amount of traffic, logs or alerts of the network infrastructure, but
also on the content side. While the term Big Data refers to the increase
in available data, this implicitly means that we must deal with problems
at a larger scale and thus hints at scalability issues in the analysis of
such data sets. Visual Analytics is an enabling technology, that offers
new ways of extracting information from Big Data through intelligent,
interactive internet and security solutions. It derives its effectiveness both
from scalable analysis algorithms, that allow processing of large data sets,
and from scalable visualizations. These visualizations take advantage of
human background knowledge and pattern detection capabilities to find
yet unknown patterns, to detect trends and to relate these findings to
a holistic view on the problems. Besides discussing the origins of Visual
Analytics, this paper presents concrete examples of how the two facets,
content and infrastructure, of the Future Internet can benefit from Visual
Analytics. In conclusion, it is the confluence of both technologies that will
open up new opportunities for businesses, e-governance and the public.
1 Introduction
We live in a world that faces a rapidly increasing amount of data. Today, in
virtually every branch of commerce and industry, within administrative and
legislative bodies, in scientific organisations and even in private households vast
amounts of data are generated. In the last four decades, we have witnessed a
steady improvement in data storage technologies as well as improvements in
the means for the creation and collection of data. Indeed, the possibilities for
the collection of data have increased at a faster rate than our ability to store
them [4]. It is little wonder that the buzzword Big Data is now omnipresent. In
most applications, data in itself has no value. It is the information contained in
the data which is relevant and valuable.
The data overload problem refers to the danger of getting lost in data, which
may be: 1. irrelevant for the current task, 2. processed in an inappropriate way, or
3. presented in an inappropriate way. In many application areas success depends
on the right information being available at the right time. The acquisition of raw
data is no longer a problem: it is the lack of methods and models that can turn
data into reliable and comprehensible information.

94 J. Davey et al.
Visual Analytics aims at turning the data overload problem into an opportu-
nity. Its goal is to make the analysis of data transparent for an analytic discourse
by combining the strengths of human and electronic data processing. Visualisa-
tion becomes the medium of a semi-automated analytical process, where humans
and machines cooperate using their distinct, complementary capabilities to ob-
tain the best results. The user has the ultimate authority in determining the
direction of the analysis. At the same time, the system provides the user with
effective means for interaction. Visual Analytics research is interdisciplinary,
combining visualisation, data mining, data management, cognition science and
other research areas. By fusing the research efforts from these fields, novel and
effective analysis tools can be developed to solve the data overload problem.
In this position paper we postulate that Visual Analytics will play a key role
in the Future Internet. We consider two facets of the Future Internet: content
and infrastructure. Both facets are characterised by vast and growing amounts
of data including the following examples: 1. Vast amounts of user-generated
content exists in private and public networks. 2. The new trend towards open
data means that ever more administrations and NGOs are making their data
available online. 3. Simulations of new architectural concepts for the Internet
generate vast amounts of data. 4. Huge repositories of network and security
data exist and are growing.
Visual Analytics researchers are already developing techniques to address the
data overload problem. Thus, we believe that these technologies can make a
significant contribution to the success of the Future Internet. With the help of
Visual Analytics, the creators and users of the Future Internet will be able to
turn data overload from a problem into an opportunity.
The rest of this article is structured as follows: Sect. 2 provides an introduction
to Visual Analytics. In the subsequent two sections, an overview of the current
and potential uses of Visual Analytics in the Future Internet is presented. In
Sect. 3 we focus on content analysis and in Sect. 4 on analysis for the improve-
ment and protection of network infrastructure. We close with a conclusion and
outlook in Sect. 5.
2 The Origins of Visual Analytics
Visual analytics emerged as the synthesis of a number of separate disciplines.

Most prominent among these were information visualization and data mining.
In this section we will briefly introduce each of these fields and then explain how
Visual Analytics developed as a new, separate research area.
2.1 Disciplines Contributing to Visual Analytics
Information Visualization (InfoVis) emerged as an independent discipline from

the scientific visualization community in the late 1990’s. Central to the formal-
ization of the field was the so-called InfoVis Pipeline shown in Fig. 1, published
in 1999 [3]. In contrast to scientific visualization, InfoVis involved the interactive
Visual Analytics 95
visualization of abstract data, i.e. data without an explicit physical or spatial

reference. As evidenced by the original InfoVis Pipeline, the first InfoVis tech-
niques were developed for tabular data. Later, techniques were developed or
extended to apply to data in other formats, such as data cubes, graphs and text
collections.
The goal of information visualization is to use images derived from data as
a means to assist users in their exploration of large data sets. Thus, it aims to
allow people to use their strongest sense, vision to think [3]. The late 1990’s and
the first years of this century saw an explosion in the number and diversity of
published visualization techniques. In the last five years, the research focus of
the InfoVis community has shifted to the evaluation of these techniques and the
development of best practices.
Fig. 1. The InfoVis Pipeline; based on the pipeline presented in [3]
Data Mining was also born in the 1990’s from the need to explore and analyse
large amounts of data. The field was first formalised in the book Knowledge
Discovery in Databases (KDD) in 1991 [17]. The so-called KDD pipeline shown
in Fig. 2 was defined in a subsequent book in 1996 [5].
In a broad sense, data mining involves the use of statistical and machine-
learning techniques to discover patterns in large data sets. Data mining tasks
include the characterization or description of data subsets, the mining of rules
describing associations or correlations, classification or regression for predictive
analysis, cluster analysis and outlier analysis [9]. Initially, these techniques were
focused on relational database management systems. However, the field has de-
veloped to include techniques for the analysis of a great variety of data sources,
including text collections, video, image and spatio-temporal data.
Fig. 2. The KDD Pipeline; based on the pipeline presented in [5]

96 J. Davey et al.
2.2 Definition of Visual Analytics

Visual Analytics was first defined by Thomas et al. as “the science of analyti-
cal reasoning facilitated by visual interactive interfaces” [19]. It emerged as an
attempt to compensate for the deficits of both data mining and information
visualization. The results of data mining algorithms are frequently difficult to
understand and often even more difficult to share with others. This lack of trans-
parency demanded a means to see the models, parameters and assumptions on
which those results were based.
Information Visualization provides techniques which allow human users to
examine abstract data with the help of visualizations. These can also be used to
expose the details of automated analysis steps. The Visual Analytics process as
proposed by Keim et al. is shown in Fig. 3 [12].
Fig. 3. The Visual Analytics Process; first presented in [12]
In 2009 and 2010 a Coordinated Action named VisMaster and funded by the
European Commission set out to establish a Visual Analytics research commu-
nity in Europe. The primary result of the project was a research roadmap entitled
Mastering the Information Age [11]. The established community has continued
its work after the project. Its main channel for dissemination and coordination
of community activities is the European Visual Analytics website.1
3 Visual Analytics for Internet Content

Despite the fact that Visual Analytics developed largely independently of Future
Internet technologies, the current trend towards visualization toolkits for the web
1
http://www.visual-analytics.eu
Visual Analytics 97
(e.g., D3 [2] or Polymaps2) suggests that visualization will play a greater role in
the near future. In addition, the number of organisations publishing their data
online is growing. As a result, new opportunities for linking and exploring these
open data sets in the Future Internet with the help of Visual Analytics arise.
3.1 Open and Public Data
During the last decade the visualization community began the creation of in-
teractive web visualizations that empowered the public to investigate open data
sets on their own. One of the most successful approaches was IBM’s ManyEyes
platform [20], which allows users to upload data, visualize it either statically
or interactively and then facilitates discussions about findings within the user
community. Besides well known charts, such as scatter plots, bar, line and pie
charts, the platform features more advanced visualizations, such as tree maps,
stacked graphs and bubble charts, and a number of textual visualizations, such
as word clouds, phrase nets and word trees. Newer web visualization tools, such
as Google’s Public Data Explorer3 and Tableau Public4 extend both the acces-
sibility of data as well as the diversity of available web visualization tools.
While web visualization tools for open data have already started to emerge,
the combination of visualization and data mining tools in Visual Analytics ap-
plications are not yet available for the web. However, we expect them to emerge
in a new wave of Visual Analytics frameworks and tools for the web.
3.2 Smart Cities
Smart Cities are characterized by their competitiveness in the areas smart econ-
omy, smart mobility, smart environment, smart people, smart living and smart
governance [8]. While strengths in each of these areas have strong links to the
historic development of cities, technological advancements such as the Future
Internet or Visual Analytics can play a role in boosting their competitiveness.
As an example, Visual Analytics applications such as the one detailed in the
study [1] can significantly empower the analysis of traffic conditions (e.g. traffic
jams) using data from GPS tags of a sample of the total vehicle population
within the city. Future Internet technologies not only play a role in the data
collection infrastructure (Internet of Things), but also in the propagation of
analysis results to commuting citizens. However, Visual Analytics is required to
turn the large and complex mobility data into useful information.
Smart governance can be enhanced through the combination of Visual Analyt-
ics and Future Internet technologies by analysing available data in the detailed
geographic context of the city. MacEachren et al. [15], for example, created a
Visual Analytics tool that takes advantage of a geo-tagged Twitter stream for
the assessment of situational awareness in application scenarios ranging from
2
http://polymaps.org/
3
http://www.google.com/publicdata/
4
http://www.tableausoftware.com/public
98 J. Davey et al.
disease monitoring, through regional planning, to political campaigning. As

demonstrated in this and the previous examples, it is the combination of Vi-
sual Analytics and Future Internet technologies that enables the advancement
of opportunities for Smart Cities.
We believe that all areas of Smart Cities can benefit from Visual Analytics
and Future Internet technologies to maintain and increase their attractiveness
for their citizens, companies and institutions.
3.3 Text and News Analysis
The Internet is full of unstructured, but often interlinked, data that could po-
tentially be valuable if processed and presented in a meaningful way. However,
issues of data processing (e.g., data quality or entity recognition) and representa-
tion (e.g., usability or scalability) turn such efforts into challenging undertakings
and only very focused approaches have so far succeeded.
Fig. 4, for example, shows a Visual Analytics system for the analysis of online
news [14] collected by the Europe Media Monitor 5 . Text clustering is used to
extract stories from the news articles and to detect merging and splitting events
in such stories. Special care is taken to minimize clutter and overlap from edge
crossing while allowing for incremental updates. Besides the main entity and
daily keywords for each story, the figure shows a list of emerging stories at the
top and a list of disappearing stories at the bottom of the screen.
Text mining can be useful for the automatic extraction of opinions or senti-
ments from user-generated content. While this data in itself is valuable, mak-
ing sense of a large collection of results can be supported using visualization as
demonstrated in the study of Kisilevich et al. [13] dealing with photo comments.
In summary, the use of Visual Analytics in the Future Internet for the analysis
of text and news data can lead to innovative web applications. However, the
unstructured nature and the linguistic intricacies of processing large but possibly
short (e.g. Twitter postings) textual data generated by a multitude of people in
several languages can impose significant challenges on the processing side.
3.4 Future Work
Currently, three projects funded by the European Commission are addressing

the challenges of Smart Governance. The projects will make use of opinion
mining and visualization technologies to draw on user-generated Internet con-
tent to inform policy-making decisions. The ePolicy project6 is focused on the
policy-making life cycle in regional planning activities. The life cycle integrates
global concerns (e.g. impacts, budget constraints and objectives) and individ-
ual perspectives (i.e. opinions, reactions extracted from the web) into the deci-
sion process, giving guidance towards better policy implementation strategies.
5
http://emm.newsbrief.eu/
6
http://www.epolicy-project.eu
Visual Analytics 99
Fig. 4. Visual Analytics for news story development. Stories are extracted from online
news articles and visualized over several days. Distinct stories about “Omar Suleiman”
and “Tahrir Square” partly merge on the 9th of February. On the 10th of February a
linked story involving the “White House” emerges.
The NOMAD project7 aims to provide politicians with tools to draw on non-
moderated sources in the Social Web for the policy appraisal. A focus will be
laid on the presentation of arguments drawn from relevant constituencies for and
against policy decisions. The FUPOL project8 aims to combine simulations of
the effects of policy decisions with information drawn from the Social Web, as
well as crowd-sourcing techniques. FUPOL will target domains such as sustain-
able development, urban planning, land use, urban segregation and migration.
While most of the interactive Visual Analytics applications currently run as
stand-alone applications, we believe that in the near future these applications
will not only take advantage of the open and public data available in the web,
but move towards client-based applications running in modern web browsers.
Furthermore, we are convinced that data linkage, text mining and modern data
management approaches will open up new opportunities for the inclusion of
Visual Analytics in Future Internet technologies. This is further supported by
the fact that streaming text data visualization (cf. [18]) is currently a hot topic
in the visualization and Visual Analytics research community.
7
http://www.nomad-project.eu
8
http://www.fupol.eu
100 J. Davey et al.
4 Visual Analytics for Network Infrastructure

The growing complexity of network infrastructure cries out for more analytical
support on both the automated side as well as on the human side. While we have
witnessed an exponential growth in networking and computing capacities, the
number of persons involved in maintaining our networks has not expanded in the
same way. Our only chance to tackle the networking issues of the Future Internet
are to either manage tasks automatically or to empower network administrators
to tackle large scale issues in a more efficient way. This section will discuss how
we can combine automated and visual approaches through Visual Analytics to
keep the Future Internet’s infrastructure alive.
4.1 Infrastructure Planning and Testing

Network infrastructure planning and testing are complex tasks. Besides historic
capacity utilization statistics, forecasting plays an important role. However, since
comprehensive interpretation of the huge volumes of data exceeds human capac-
ities, meaningful abstractions and a focus on specific sub-problems are necessary
to master the network’s complexity. Visual Analytics builds on human percep-
tual capabilities to spot interesting patterns and automatic methods to deal
with large scale data and thus enables interpretation at a higher level of detail.
Furthermore, interaction methods extend Visual Analytics methods and enable
exploratory analysis tasks.
Hierarchical Network Maps [16] are one example of how visualization can
facilitate the interpretation of network capacities. In particular, this technique
uses a hierarchy of continents, countries, autonomous systems and IP prefixes
to render a TreeMap [10] of the internet. Coloring can then be used to match
traffic load onto rectangles and interaction facilitates drill-down along the levels
of the hierarchy for chosen regions.
4.2 Network Security

Today, signature-based and anomaly-based intrusion detection are considered
state-of-the-art in network security. However, fine-tuning parameters and analy-
sing the output of these intrusion detection methods can be complex, tedious,
and even impossible when done manually. In general, systems become more and
more sophisticated and make decisions on their own up to a certain degree.
However, as soon as unforeseen events occur, system administrators or security
experts have to intervene to handle the situation. While network monitoring
and security have profited a lot from automatic detection methods in recent
years, visual approaches foster a better understanding of the complex informa-
tion through interactive visualization and therefore have a lot of potential to
complement the former approaches.
By means of the Visual Analytics application NFlowVis [7] we demonstrate
in this section how the combination of automatic and visual analysis can help
security experts to derive more meaning out of the vast amount of security events
Visual Analytics 101
Fig. 5. Analysis of a distributed network attack on the SSH service of a university

network on May 11, 2008 using NFlowVis [7]. The circles on the outside represent
hosts that attack the squared hosts in the internal network. A clustering algorithm
ensures that related attackers (see top) are positioned next to each other.
and traffic data which is characteristic of the field. In particular, we use traffic
patterns which are common for signature-based intrusion detection systems and
one day of network traffic statistics (NetFlows) from the main gateway of a
medium sized university, which amounts to approximately 10 GB of raw data.
Fig. 5 shows the visual output of an analysis with NFlowVis. After having se-
lected suspicious hosts from the intrusion detection system, their network traffic
to all hosts in the internal network is retrieved from a database and visualized.
While automatic intrusion detection systems output many alerts in a large net-
work, the visualization supports the analyst in the difficult task of correlating
these alerts with each other and setting them into context. In this particular
case, we chose an SSH traffic pattern and visualized a number of external hosts
matching this traffic pattern.
Before visualizing the information, the system first clusters the external hosts
(potential attackers) and then places them on the nearest border in such a way
that a) hosts with similar traffic patterns appear next to each other and b)
preferably short splines are drawn to connect the dots of the external hosts
and the rectangles representing their internal communication partners. Color
encodes the first byte of the IP address of the external host in such a way that
attackers from nearby network prefixes are drawn in a similar color. This helps
102 J. Davey et al.
to judge whether the attack is conducted from a particular network or from hosts
distributed all over the Internet.
Drawing straight connecting lines results in a lot of visual clutter. To reduce
this clutter, the lines are grouped by exploiting the structure of the underlying
hierarchical visualization of the /24 prefixes. As a result, the analyst can easily
identify the pattern of the distributed attack on the upper right of Fig. 5, which
details a number of external hosts targeting the same subset of internal hosts
in the university network. A more detailed analysis revealed that all attacking
hosts contacted 47 hosts and thereby consciously avoided a common threshold of
an automatic intrusion detection system. The visual output furthermore shows
scanning activity of individual hosts on the lower left and top right of Fig. 5. We
assume that scanning activity first identified candidate victims in the network
and that the botnet then used this information to target this subset of hosts in
the university network, since the number of attacked hosts per subnet varies.
Currently, the VIS-SENSE project9 , funded by the European Commission, is
applying Visual Analytics techniques to large, network-security-related data sets.
The project focuses on the strategic analyses of spam, malware and malicious
websites. In addition, the misuse of the Border Gateway Protocol for criminal
activities will be analysed.
4.3 Real-Time Monitoring

Modern services heavily rely on the availability of the network and server infras-
tructure to comply with the strict service level agreements of business users and
consumers. However, defining a valid state for all components of the network is
not possible due to the high number of complexities and inter-dependencies of
all involved systems. Modern monitoring approaches therefore often produce ei-
ther too many or too few alerts, which makes manual analysis close to real-time
almost impossible.
In this case Visual Analytics can bridge the gap between the complexity of
the data and the human understanding and thus speed-up both investigation of
failures and system recovery operations. The work in [6], for example, details a
Visual Analytics system for the analysis of system log events in real-time. With
peaks of up to 425,000 events per hour, the interactive time-line visualization and
the geographic map interface highlight events according to a scoring model and
enable the detection of unusual activity, such as remote accesses from uncommon
sources or bursts of critical events on servers.
4.4 Future Work

While this section detailed some exemplary uses of Visual Analytics for planning,
monitoring and securing network infrastructure, many tasks in this wide field are
still conducted without any visual or computational support. We therefore see a
lot of potential for research that connects the still largely independent fields of
Visual Analytics and the Future Internet.
9
http://www.vis-sense.eu
Visual Analytics 103
5 Conclusion
In this article we presented an introduction to Visual Analytics and its relevance
for the Future Internet. We considered the two facets content and infrastructure.
Both facets are characterized by a vast and growing amount of data.
With respect to content in the Future Internet, we have shown that emerging
data visualization platforms for the web derive their value from the relevance
of the data that is analysed with them. Since more and more open and public
data becomes available every day, it is only a matter of time before existing
visualization platforms hit scalability limits – due to the data overload problems
at hand – and need to include automated data analysis functionality. While the
analysis of the abundance of text and news available in modern media like Twit-
ter imposes significant challenges, working on these problems can have drastic
effects on the development of countries, regions and smart cities. We are thus
convinced that targeted research in Visual Analytics can revolutionize the way
in which we interact with content in the Future Internet.
Besides its potential for content, Visual Analytics can play an important role
in the network infrastructure of the Future Internet. Due to the amount of data
available from networking devices, the inherent complexity of the network and
the need to immediately react to failures or attacks, visual and computational
support for tasks in this domain can significantly improve infrastructure plan-
ning and testing, as well as network monitoring and security. We conclude that
strengthening the connection between Visual Analytics and the Future Internet
will enable us to build a more secure, reliable and scalable network.
The examples presented show how Visual Analytics is already contributing
solutions to the data overload problem in the Future Internet. Thus we are
convinced that the confluence of both technologies has enormous potential for
use in the business, administrative and private spheres.
credited.
References
1. Andrienko, G.L., Andrienko, N.V., Hurter, C., Rinzivillo, S., Wrobel, S.: From
movement tracks through events to places: Extracting and characterizing signifi-
cant places from mobility data. In: IEEE Conference on Visual Analytics Science
and Technology (VAST 2011), pp. 161–170 (2011)
2. Bostock, M., Ogievetsky, V., Heer, J.: D3: Data-driven documents. IEEE Transac-
tions on Visualization and Computer Graphics 17(12), 2301–2309 (2011)
3. Card, S.K., Mackinlay, J.D., Shneiderman, B.: Readings in information visual-
ization: using vision to think. Morgan Kaufmann Publishers Inc., San Francisco
(1999)
4. Cukier, K.: Data, data everywhere: A special report on managing information. The
Economist 1(1), 14 (2010)
104 J. Davey et al.
5. Fayyad, U.M., Piatetsky-Shapiro, G., Smyth, P., Uthurusamy, R. (eds.): Advances

in Knowledge Discovery and Data Mining. AAAI Press (1996)
6. Fischer, F., Mansmann, F., Keim, D.A.: Real-Time Visual Analytics for Event Data
Streams. In: Proceedings of the 2012 ACM Symposium on Applied Computing,
SAC 2012. ACM (2012)
7. Fischer, F., Mansmann, F., Keim, D.A., Pietzko, S., Waldvogel, M.: Large-Scale
Network Monitoring for Visual Analysis of Attacks. In: Goodall, J.R., Conti, G.,
Ma, K.-L. (eds.) VizSec 2008. LNCS, vol. 5210, pp. 111–118. Springer, Heidelberg
(2008)
8. Giffinger, R., Fertner, C., Kramar, H., Kalasek, R., Pichler-Milanovic, N.,
Meijers, E.: Smart cities ranking of european medium-sized cities (2009),
http://www.smart-cities.eu/ (retrieved January 20, 2012)
9. Han, J., Kamber, M., Pei, J.: Data Mining: Concepts and Techniques., 3rd edn.
Morgan Kaufmann Series in Data Management Systems. Morgan Kaufmann Pub-
lishers Inc., Waltham (2012)
10. Johnson, B., Shneiderman, B.: Tree-maps: A space-filling approach to the visual-
ization of hierarchical information structures. In: Proc. IEEE Conference on Visu-
alization, pp. 284–291. IEEE (1991)
11. Keim, D.A., Kohlhammer, J., Ellis, G., Mansmann, F. (eds.): Mastering The
Information Age - Solving Problems with Visual Analytics. Eurographics (2010),
http://www.vismaster.eu/wp-content/uploads/2010/11/VisMaster-book-
lowres.pdf
12. Keim, D.A., Mansmann, F., Schneidewind, J., Thomas, J., Ziegler, H.: Visual An-
alytics: Scope and Challenges. In: Simoff, S.J., Böhlen, M.H., Mazeika, A. (eds.)
Visual Data Mining. LNCS, vol. 4404, pp. 76–90. Springer, Heidelberg (2008)
13. Kisilevich, S., Rohrdantz, C., Keim, D.A.: Beautiful picture of an ugly place. Ex-
ploring photo collections using opinion and sentiment analysis of user comments.
In: Computational Linguistics & Applications (CLA 2010), pp. 419–428 (October
2010)
14. Krstajic, M., Najm-Araghi, M., Mansmann, F., Keim, D.: Incremental Visual Text
Analytics of News Story Development. In: Proceedings of Conference on Visual-
ization and Data Analysis, VDA 2012 (2012)
15. MacEachren, A.M., Jaiswal, A.R., Robinson, A.C., Pezanowski, S., Savelyev, A.,
Mitra, P., Zhang, X., Blanford, J.: Senseplace2: Geotwitter analytics support for
situational awareness. In: IEEE Conference on Visual Analytics Science and Tech-
nology (VAST 2011), pp. 181–190 (2011)
16. Mansmann, F., Keim, D.A., North, S.C., Rexroad, B., Sheleheda, D.: Visual Anal-
ysis of Network Traffic for Resource Planning, Interactive Monitoring, and Inter-
pretation of Security Threats. IEEE Transactions on Visualization and Computer
Graphics 13(6) (2007)
17. Piatetsky-Shapiro, G., Frawley, W.J. (eds.): Knowledge Discovery in Databases.
MIT Press (1991)
18. Rohrdantz, C., Oelke, D., Krstajic, M., Fischer, F.: Real-Time Visualization of
Streaming Text Data: Tasks and Challenges. In: Workshop on Interactive Visual
Text Analytics for Decision-Making at the IEEE VisWeek 2011 (2011)
19. Thomas, J.J., Cook, K.A. (eds.): Illuminating the Path: the Research and Devel-
opment Agenda for Visual Analytics. IEEE CS Press (2005)
20. Viegas, F., Wattenberg, M., Van Ham, F., Kriss, J., McKeon, M.: Manyeyes: a
site for visualization at internet scale. IEEE Transactions on Visualization and
Computer Graphics 13(6), 1121–1128 (2007)
Towards a Trustworthy Service Marketplace
for the Future Internet
Francesco Di Cerbo , Michele Bezzi, Samuel Paul Kaluvuri,

Antonino Sabetta, Slim Trabelsi, and Volkmar Lotz
SAP Research
805, Av. du Docteur Maurice Donat
06250 Mougins, France
[email protected]
Abstract. Digital economy is moving towards offering advanced busi-

ness services, integrated into different applications and consumed from
heterogeneous devices. Considering the success of actual software market-
places, it is possible to foresee that Service Marketplaces (SM) will play
a key role for the future Internet of Services. At present, on all offered
software, marketplace operators define requirements that are common,
and are validated before admitting them. However, the requirements,
the validation process, and its results are not completely evident to the
service consumers, resulting in a significant shortcoming especially with
respect to security characteristics. In addition, having common security
requirements for all services and applications makes the validation possi-
bly inadequate to address the specific requirements that consumers may
have.
In order to address these points, we propose the concept of a trust-
worthy service marketplace for the upcoming Internet of Services, where
the security characteristics of services are certified and treated as first-
class entities, represented in a machine-processable format. This allows
service consumers – either human end-users or computer agents – to rea-
son about these security features and to match them with their specific
security requirements.
Keywords: Security, Trustworthiness, Trust, Service Marketplace.
1 Introduction
The marketplace metaphor is increasingly pervasive in today’s digital economy.
A software marketplace is a virtual place, where software providers can adver-
tise their “apps” or services, and customers can browse and buy them; software
marketplaces offer a centralized application distribution mechanism that reaches
immediately many potential customers, all over the world. Marketplaces dedi-
cated to specific devices or operating environments are nowadays proliferating
and they represent a valuable business opportunity for software vendors. In many

Corresponding author.

106 F. Di Cerbo et al.
cases, like for the Apple Store[4], Windows Marketplace[17], or the Amazon Kin-
dle Store[1], they are evolving to become gateways to entire ecosystems, with a
potential audience of millions.
Similarly to apps, services can leverage on the marketplace distribution chan-
nel. Services relieve consumers from the burden of acquiring and managing their
own operational infrastructure, on top of the benefits of component-based soft-
ware [24]. Nowadays, following the SaaS (Software-as-a-Service) model, services
are more and more commonly consumed as “commoditized” pieces of functional-
ity, and are extensively adopted as a means to increase flexibility and to optimise
IT expenditure.
The very nature of the model aims at simplifying software consumption by
insulating the consumers from the complexity related to deployment, operation,
and management of the software. However, in the process, important informa-
tion about the quality of the software are not evidently reported to consumers,
raising a relevant challenge with regard to the trust of the consumers on software
providers. In addition, the centralized nature of most of software marketplaces
results in “one size fits all” security checks, which are not appropriate for many
security-critical applications, typically characterized by domain-specific require-
ments.
We believe that addressing these challenges is key to the success of the fu-
ture Internet of Services, especially with respect to services that are considered
highly valuable, sensitive critical or in the context of serious applications. Two
key factors can contribute to that: the availability of a more detailed description
of the security features of services and the possibility to include some addi-
tional guarantees on the quality of security mechanisms provided by established,
domain-specific security experts (security certifications).
It is crucial that this information be provided to service consumers (human
or software agent) in a machine-readable form such that they can check directly
and just-in-time what specific security features are provided, what assurance
they can get from a software product and how this assurance is provided. In this
paper we introduce the concept of a trustworthy Service Marketplace (SM) that
is suitable for hosting a larger class of security- and business-critical services,
and service compositions for both businesses and end-users.
The remainder of this paper is organised as follows: Section 2 contains an
overview on the state of the art in software marketplaces, Section 3 details the
major challenges to be addressed towards a trustworthy SM, with particular
attention to the limitation of current security certification schemes. Sections 4
presents our approach to tackle these challenges, while Section 5 illustrates the
vision of a trustworthy SM. Finally, Section 6 concludes the chapter.
2 State of the Art
Before introducing the concept for trustworthy SMs, we analyse the state-of-
the art in software marketplaces, and their relevant security checks. We focus
mainly on mobile software markets, as they provide a large user base and are the
Towards a Trustworthy Service Marketplace for the Future Internet 107
subject of many studies. This section is composed of two parts: we review the
main marketplace approaches to security, then we focus on the security checks
performed when an application/service has to be admitted in the marketplace
(the vetting process).
2.1 Software Marketplaces

Security and trust play a major role for software market customers, for pro-
fessional service users but also in the mobile device application consumption,
due to the high sensitivity of information typically kept in mobile phones. In-
terestingly enough, in many markets, like the Apple App Store, security is not
guaranteed [5], even if sometimes users’ perception is different: what is provided
is the availability of prompt security procedures, like the “kill switch” option,
i.e., the automatic removal of any application, instructed for instance by Apple
or Google, on all their produced mobile phones, without user intervention.
Marketplace operators can adopt different approaches to deal with security
while delivering applications to end users: in particular, Barrera and Van
Oorschot[5] propose three categories, “Walled garden”, “Guardian” and “User
control”; they range from a rigorous assessment of any applications on the mar-
ket, to a completely open model, where security checks are upon user’s respon-
sibility. They also propose a classification of vetting tests for applications to
be advertised on a (mobile software) marketplace. The seven categories are:
“smoke tests”, “hidden-API checks”, “functionality checks”, “intellectual prop-
erty, liability and terms-of-service checks”, “UI checks”, “bandwidth checks”,
and “security checks”.
In many software markets, the vetting processes are not described in details,
we will discuss this aspect in the following Section 2.2.
Researchers have different opinions with respect to the role that software mar-
ketplaces can play in improving trustworthiness, and in bringing security to end-
users, by means of their security assessments. Some authors stressed the difficulty
to define a common concept of “security” for all users, given the multitude of dif-
ferent security requirements, according to contexts, users, and applications[14].
In the same paper, McDaniel and Enck argue about the possibility to introduce
automated tests at application publishing phase, to check configuration settings,
binaries and source code. The results should be then pushed to end users, allow-
ing them to take the final decision about installing an application or not, based
on their own security definition and requirements.
Other authors underline the role that central application repositories can have
in filtering out malware applications. In particular, Gilbert et al. [10] analysed
the benefits provided by a dynamic-analysis security validation tool that could
be integrated in the software market approval process, but also scanning pe-
riodically the software market applications. The authors claim that such tool
could be useful for specific purposes, like for instance for protecting the end
users from privacy threats. Lastly, there is a certain emphasis given by some au-
thors on publicly disclose the obtained results of software market assessments.
For McDaniel and Enck, but also for Gilbert et al., the public availability of the
evaluation process can contribute to a more conscious use of technology by end

users; especially with respect to sensible information leaking, letting them to be
aware of risks they can be exposed to.
2.2 Vetting Process

We present in Table 1 a number of relevant marketplaces, together with their
publicly disclosed security assessment criteria.
Salesforce releases a customer relationship management (CRM) system on the
cloud that has a number of companion tools.
It permits third-parties to publish and advertise their applications (or exten-
sions to existing Salesforce applications) that can operate on customers’ data and
information, on a specific marketplace with defined security review policies [23].
Google Apps Market is a store where third-parties can advertise complementary
services for Google Apps services. Google explicitly inform its customers that no
security checks are conducted on advertised applications [11]. Windows Azure
Market is the official marketplace for Windows Azure (Platform-as-a-Service).
Third parties can advertise their services, that apparently are not verified by
Microsoft [16]. Existing marketplaces adopt the previously-described “User Con-
trol” approach. App Store and iOS, instead, can be seen as examples of “Walled
Garden”, meaning that anything that runs on served mobile devices must be
explicitly approved by Apple. The app review process is not publicly disclosed;
in a response to a FCC request in 2009, Apple disclosed some information[3],
that are contained in Table 1. Microsoft offers Windows Marketplace [17] to
users of its Windows Mobile OS. Application publishing and review process is
documented in MSDN[15], the reference guide for any development effort with
Microsoft technologies. Also Nokia has a specific certification process for pub-
lishing apps on its market [19], the Nokia Store[20]; nevertheless, newer Nokia’s
Windows mobile phones should follow Microsoft guidelines. RIM’s App World
is the reference software market for BlackBerry devices. Almost no public infor-
mation on security assessment could be found, except those contained in [22].
In summary, where applicable, none of the above marketplaces discloses:
– the details of its security assessments, or
– the results of the vetting process for each applications.
This means that users have to cope with a “one-size-fits-all” definition of security,
like in the majority of cases, having no option but to trust blindly marketplaces’
procedures; or they have to face the absence of security assessments, having no
option but to trust third-parties.
3 Challenges for Trustworthy Service Marketplaces

As discussed in the previous section, most marketplace operators enforce some
sort of review and evaluation processes on applications before they are admitted
to their marketplace. Security evaluation may involve security experts from the
Table 1. Security Features Of Existing Software Markets. Information marked with

’*’ are not completely publicly disclosed by providers.
Market name Code Architectural Hands-on as- Periodic secu- Application

reviews review sessment rity review Removal
Salesforce AppExchange No Yes Yes Yes Yes
Google AppsMarket No No No No Maybe
Windows Azure Marketplace No* No* No* No* Maybe
App Store No* No* No* No* Yes
Android Market No No No No Yes
Windows Mobile Marketplace No Yes No Yes Yes
Nokia Store No No No No Yes
BlackBerry App World No* No* No* No* Yes
marketplace operator and/or a third party security organization, who approve

an application if it satisfies the security requirements defined by the marketplace
operator.
The admission process of the marketplace compels the application providers
to develop applications that address the security criterion specified by the mar-
ketplace operator.
However, this approach does not scale for different software provisioning sce-
narios.
Though the vetting processes increase the trust of the consumer on the se-
curity of the applications offered through the marketplaces, especially in the
vision of a service marketplace, there are important problems that need to be
addressed:
1. There is no information about the outcome of an evaluation available for
the consumer, and the evaluation process is not disclosed in detail. Hence,
trust in the secure operation of an application can only be built based on
the reputation of the marketplace operator.
2. Consumers have specific security requirements for applications based on the
operating domain and/or usage of the applications. However, marketplace
operators have limited application- and domain-specific knowledge which is
essential to perform any meaningful and effective evaluation on the security
of the application, in a way that addresses the specific security requirements
of consumers.
3. Current admission processes require the marketplace operators to own/con-
trol the execution environment of the applications, which is true for most
of the current marketplaces. However this may not be the case in future
marketplaces, especially in service marketplaces.
4. Admission checks cannot provide end-to-end security assurance for an appli-
cation, especially when applications consume external services.
This means that the security requirements for a service significantly depend on
the application domain, the application context, and the business context (in-
tended usage). Hence, the security properties that a service provides should be
evaluated and consequently certified by specialized entities that have the re-
quired domain- and application-specific knowledge. The lack of assurance on the
security of services is one of the key reasons of the trust deficit of consumers on
such services [18]. Security certification of services can bridge this trust deficit
by providing the required assurance on service’s security. Though current se-
curity certification schemes are successful in providing assurance in monolithic
software systems, they suffer from severe limitations when applied in a service
environment due to economic and technological factors.
In addition, the stakeholders, the consumption models of current certification
schemes are modelled for monolithic software and hence current schemes are
inadequate to provide the security assurance in a service environment.
Some of the shortcomings of current certification schemes have conceptual rea-
sons. Schemes such as Common Criteria are intentionally designed to be flexible
and generic, in order to be able to certify different products ranging from software,
firmware to hardware [25]. However this prevents these schemes to be prescriptive
and so comparing certificates of different products becomes complex.
In addition, current certification schemes are structured in a manner that they
cater to software provisioning paradigms where the consumer has control over
the operation and execution of the product. However, in the service-oriented
computing paradigm, the consumer does not have any control over the opera-
tional environment nor on the execution environment.
Another limitation is the application of current schemes in practice is a very
expensive and time consuming process, often requiring years even for medium-
level security assurance [25]. This is a major obstacle for services, where time-
to-market can be a critical factor for the success of the service. Schemes such
as Common Criteria allow a lightweight certification, but they lead to very low
assurance. Also the evaluation is focused more on the accompanying documen-
tation (Architecture, Design, Process related etc.,) or on the security processes
followed, rather than the actual implementation of the product, especially at
lower assurance levels.
The certification process, and results of the evaluation are captured in a hu-
man readable form that do not allow automated reasoning and processing to be
performed. This is one of the major challenges that hampers the usage of current
security certification schemes to service marketplaces where the security require-
ments of the consumers must be easily matched with the security properties of
the services.
4 Building Blocks of Service Marketplaces of the Future

Internet
4.1 Security Certification for Services: Assert4Soa
Current certification schemes have to tackle new challenges when approaching
Internet of Services (IoS), for expressing, evaluating and certifying security prop-
erties for service-oriented applications. Therefore, novel models, techniques and
tools are much needed; the Assert4Soa project aims at providing answers to
these requests, defining a specific methodology as well as companion artefacts
and tools [2,6].
Similarly to current security certification schemes, in Assert4Soa the as-

sessment of the security properties of a service is performed by an independent
third party (certification authority), who issues a corresponding signed assess-
ment (Assert ), bound to the service. The certification of a security property in
an Assert is based on either a formal proof or on service testing that has been
carried out before the certificate is issued. These formal proofs and tests must
have been carried by the dedicated evaluation entity that has been accredited by
the certification authorities. The Assert4Soa certification process will be semi-
automated by using extensive tool support, as opposed to current certification
schemes that depend heavily on manual effort.
A core feature of the Assert4Soa approach is a language, designed to express
the security properties of a service as machine-readable, digitally signed state-
ments (asserts), as opposed to existing security certificates that are expressed
in a human readable form. The language allows the security features of a service
to be represented at different levels of granularity ranging from abstract security
properties to actual security functionalities that are implemented in the service.
This is done in order to cater to the specific needs of different types of consumers
that can range from users who have limited knowledge of service security to se-
curity experts of organizations who have specific requirements on the security
functionalities of a service. The language also enables the representation of an
abstract model of the service as part of the target of evaluation. This not only
provides a description of the service to the consumers, but also serves to mitigate
the concerns of the consumers on the lack of transparency of services.
In addition to the certified security properties, the language allows the repre-
sentation of the information about the certification authority that has issued the
certificate as well as the evidence that underpins the certified properties, i.e., the
test suites or formal proofs used to evaluate the service. Hence, Asserts provide
comprehensive descriptions of the security properties of the service.
Another important feature of the Assert4Soa project is the service-discovery
framework. The service discovery framework provides consumers a query lan-
guage through which they can express the functional and security requirements
on the services. The query language allows the consumers to express the security
properties at different levels of granularities as well as their preferences on the
type of evidences for those security properties. The discovery engine, which is at
the core of the service discovery framework, processes consumers requirements
and performs matchmaking on the functional and security requirements using
the functional and security matchmakers.
4.2 Component: USDL-SEC

Services published in marketplaces should be described in a manner that en-
ables their discovery based on not only the functional requirements but also the
security requirements of the consumer. However, the current description lan-
guages are not capable of describing the security properties of services. Though,
some languages such as OWL-S [13] recommend using existing standards such
as WS-Security [8], SAML [7] to describe security-relevant properties, they do
not provide a comprehensive specification. In order to overcome this limitation,

we propose Usdl-Sec, a new security specification model, that describes the
security properties of services. This specification can extend existing service de-
scription languages such as Usdl [21]. Service providers can use this specification
to describe the security features of their services, and thus to support users in
finding adequate alternatives to fulfil their needs.
The Usdl-Sec model described here is currently being developed in the con-
text of the EU-Funded Fi-Ware1 project. This model is globally organised in
three main layers:
– Security Topic: This is a high level representation of the security feature
of a service.
– Security Solution: This is a security mechanism that contributes towards
satisfying a particular security topic.
– Security Technology: It refers to the technical implementations of the
security solutions.
This three-layered model is materialized into a more concrete description model,
depicted in Figure 1.
The model in composed by the following elements:
– Security Profile: the root node of the model and the entry point from
Usdl to Usdl-Sec. This node should appear as a pointer element of Usdl
to the security properties of the service. This pointer can assume two dif-
ferent values, reflecting the categorization expressed in the previous section
“ Usdl-Sec target”: Security service, that refers to the Security-as-a-service
paradigm, or service with security features, indicating that the service is a
generic service with security capabilities.
– Security Goal: the security goal refers to the highest abstraction layer
referring to a security topic. It can take the values of the most well known
security concepts like Anonymity, Confidentiality, Privacy, Authentication
etc. This list is defined using a security ontology ([12]).
– Security Mechanism: is a set of security solutions that can achieve a secu-
rity goal. These mechanisms are theoretical solutions that answer to specific
security requirements like Access control, Cryptography, Obligations, etc.
These solutions can be applied under three realization levels: The network
level, the application level, and the service level.
– Security Technology: is a set of concrete implementations and tools that
realizes the security mechanisms. Like for example the encryption on the
network level is implemented by IPSec [9].
As a use case example, the Data Handling GE service being developed in the
Fi-Ware project is described using Usdl-Sec, as shown in Listing 1.1. This is a
security service that protects sensitive data, by associating to each data transfer
a specific privacy policy, and by enforcing its application. This service is assumed
to be described in USDL for its business-related features. The Usdl-Sec security
1
www.fi-ware.eu
profile illustrates the security goals of the service (Privacy and Authorization);
it also indicates the security mechanisms and technologies adopted to meet the
security goals (Obligation and PPL Language in one case, AccessControl and
XACML in the other).
<<enumeration>>
SecurityGoal
<<enumeration>> 1 1..*
SecurityProfile Accountability
SecurityFeaturesType
Anonymity usdl:ServiceDescription
SecurityService goals Authentication 1 1
ServiceWithSecurityFeatures Confidentiality
etc
mechanisms referenceFeature 1..*

1
<<enumeration>> SecurityMechanism
SecuityRealizationType 1 1 SecurityTechnology
Network 1 1..*
App
Service referenceMechanism technologies
AccessControl Cryptography KeyManagement Monitoring AES Blowfish XACML
Fig. 1. USDL-SEC Specification Model
Listing 1.1. Draft for Data Handling GE in Fi-Ware platform
<#usdlSecDHGESecurityProfile> a sec : S ecu r i t y P r o f i l e ;

dc : t i t l e ” S e c u r i t y p r o f i l e o f Data H a n d l i n g GE” ;
sec : providesSecurityFeature [
a sec : SecurityFeature ;
sec : hasRealizationLevel sec : Service ;
s e c : h a sS e c u r i t y M e c h a n i sm
[
a sec : Obligation ;
s e c : h a s I m p l e m e n t a t i o n s e c : PPL
],
[ a sec : AccessControl ;
s e c : h a s I m p l e m e n t a t i o n s e c :XACML ]
];
s e c : h a s S e c u r i t y G o a l s e c : Privacy , s e c : A u t h o r i z a t i o n .
: PPLService a u sd l : S e r v i c e ;
s e c : h a s S e c u r i t y P r o f i l e <#usdlSecDHGESecurityProfile>.
5 Towards Trustworthy Service Marketplaces

Consider the scenario of a service consumer, who uses a SM to discover a service
providing file storage functionality, in addition the consumer also has a secu-
rity requirement that the file should be stored in a confidential manner. Now
let us assume that there exits a service s, that provides confidential file stor-
age. In the current SMs, the consumer cannot discover this service, as service
discovery based on security properties of services is not supported. Even if the

consumer is able to discover the service s, there is still a lack of assurance that
the security property of the service is indeed implemented correctly. We aim to
overcome these limitations through the concept of Trustworthy Service Market-
place (TSM).
Our vision for TSM combines the description of service security features with
supporting security certificates. While Usdl-Sec allows the representation of
the security features, the Asserts (Security Certificates) provide assurance to
the consumers on the security features of services by providing evidences used to
evaluate the services. These two approaches complement each other and together
contribute towards increasing the trust of the consumer on the services offered
through the marketplace.
The service providers should describe the security features implemented in
the service using the Usdl-Sec specification model before publishing them on
the SM. They should describe the security topic, the security solutions and the
security technologies implemented in the service. Though description of security
features enables consumers to discover services that meet their security require-
ments, there is a lack of assurance that the security features are actually present
and implemented correctly. In order to provide this assurance, service providers
can obtain a security certification that evaluates the service thoroughly by using
test suites or formal models.
The SM operator should use an advanced query language, that can be used
by the consumers to express not only their business requirements, functional
requirements but also their security requirements, assurance requirements and
preferences. The Usdl query language developed in the Fi-Ware project allows
the consumers to express their business, technical and functional requirements
among others. The query language developed in the Assert4Soa project can
be used for expressing the specific security, assurance requirements, and security
preferences on the services. In this manner, a wide range of requirements can be
used for querying the SM.
The traditional service discovery engines of the SMs should be augmented to
use the Usdl-Sec Engine and the Assert Service Discovery (ASD) framework.
The Usdl-Sec engine matches the requirements of the consumer with the secu-
rity features of the services based on their Usdl-Sec descriptions. The (ASD)
framework allows the SM to discover certified services based on their security and
assurance requirements and present them to the consumer. The ASD framework
employs a matchmaking system that ranks services based on their degree of fit to
the consumer’s requirements. Though at a high level, there is an apparent over-
lap in the functionalities of the Usdl-Sec engine and the ASD Framework, the
functionalities complement each other in practise, where the Usdl-Sec engine
performs matchmaking on the abstract security requirements with service secu-
rity descriptions, and the ASD framework performs matchmaking on the refined
security requirements with certified properties of services along with their evi-
dences. Together they provide a ranked list of services (recommendations) that
match the business, functional, technical, security, and assurance requirements

of the consumers.
In addition to using the Usdl-Sec and Asserts the SM operator could employ
a vetting process, however the processes, the results of the vetting process must
be made transparent to the consumer. The SM operators could also prescribe
the services to comply with a standard Usdl-Sec profile, accompanied by a
security certification performed by independent Certification Authorities.
In the scenario mentioned above, if the consumer uses the TSM, he would
not only be able to discover the service s based on the functional and security
requirements, but also have assurance that the security requirements are actually
met by the service.
6 Conclusions
Trustworthy Service Marketplaces can represent a key factor for opening new
market perspectives for the future Internet of Services, especially with respect
to sensitive, critical services and service composition. Trustworthy SMs will serve
all their stakeholders with advanced and more secure services, as well as with
transparent and evidence-based vetting processes. They will enable refined ser-
vice discovery operations in marketplaces, also according to specific security re-
quirements. Candidate services shall be then presented to users, along with their
security certificates and evidences. In this way, a customer could evaluate each
alternative according to her specific operational scenario. Trustworthy SMs could
set certain security thresholds, such that a minimal security standard will have to
be met by any of their advertised element. To sustain this vision, new technolo-
gies and standards are in development: digitally consumable service descriptions,
covering business , technical, security and contextual aspects (Usdl/Usdl-Sec
in Fi-Ware); new assessment and certification methodologies, as well as digitally
consumable certificates (Assert4Soa). Relying on assumptions and constraints
expressed, more functionalities will come, like for instance a support for secure
service compositions, through analysing security requirements and prerequisites
of services, and secure deployment of services. We believe that trustworthy SMs
can increase the trust and confidence in Internet-based systems, thus enabling
even more sensitive operations to take place, in a secure, reliable and effective
way.
Acknowledgements. This work is partially supported by projects Fi-Ware

(FP7-2011-ICT-FI - www.fi-ware.eu) and Assert4Soa (Grant No. 257351 -
www.assert4soa.eu).
credited.
References
1. Amazon. Kindle, http://www.amazon.com/kindle-store-ebooks-newspapers-
blogs
2. Anisetti, M., Ardagna, C.A., Guida, F., Gürgens, S., Lotz, V., Maña, A., Pandolfo,
C., Pazzaglia, J.-C.R., Pujol, G., Spanoudakis, G.: ASSERT4SOA: Toward Secu-
rity Certification of Service-Oriented Applications. In: Meersman, R., Dillon, T.,
Herrero, P. (eds.) OTM 2010. LNCS, vol. 6428, pp. 38–40. Springer, Heidelberg
(2010)
3. Apple inc. FCCs answers, http://www.apple.com/hotnews/apple-answers-fcc-
questions/
4. Apple inc. Official apple online store, http://store.apple.com/us
5. Barrera, D., van Oorschot, P.: Secure software installation on smartphones. IEEE
Security & Privacy 99, 1 (2010)
6. Bezzi, M., Sabetta, A., Spanoudakis, G.: An architecture for certification-aware
service discovery. In: Proc. of IWSSC (co-located with NSS 2011) (2011)
7. Cantor, S., Kemp, I., Philpott, N., Maler, E.: Assertions and protocols for the oasis
security assertion markup language. OASIS Standard (March 2005)
8. O. W. S. S. Committee. OASIS web services security (WSS) TC OASIS,
http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=wss
9. Doraswamy, N., Harkins, D.: IPSec: the new security standard for the Internet,
intranets, and virtual private networks. Prentice Hall (2003)
10. Gilbert, P., Chun, B., Cox, L., Jung, J.: Vision: automated security validation of
mobile apps at app markets. In: Proceedings of the Second International Workshop
on Mobile Cloud Computing and Services, pp. 21–26 (2011)
11. Google inc. Evaluate a marketplace app’s security, https://support.google.com
12. Herzog, A., Shahmehri, N., Duma, C.: An ontology of information security. Inter-
national Journal of Information Security 1(4), 1–23 (2007)
13. Martin, D., Burstein, M., Hobbs, J., Lassila, O., McDermott, D., McIlraith, S.,
Narayanan, S., Paolucci, M., Parsia, B., Payne, T., et al.: OWL-S: semantic markup
for web services. W3C Member Submission 22, 200704 (2004)
14. McDaniel, P., Enck, W.: Not so great expectations: Why application markets
haven’t failed security. IEEE Security & Privacy 8(5), 76–78 (2010)
15. Microsoft inc. Market, http://msdn.microsoft.com/en-us/library/gg490776.
aspx
16. Microsoft inc. Windows azure: Terms of use, https://datamarket.azure.com/
terms
17. Microsoft inc. Windows marketplace, http://www.windowsphone.com/marketplace
18. Nasuni. Security and control are greatest concerns preventing enterprises from
adopting cloud storage, http://www.nasuni.com/news/press_releases/
19. Nokia. Nokia ovi store content guidelines, http://support.publish.nokia.com
20. Nokia. Packaging and signing, http://www.developer.nokia.com/
21. Pedrinaci, C., Leidig, T.: Linked-USDL, http://linked-usdl.org/ns/usdl-core
22. RIM inc. BlackBerry app world, http://us.blackberry.com/developers/
appworld/
23. Salesforce. Security review, http://wiki.developerforce.com/page/Security
Review
24. Szyperski, C., Gruntz, D., Murer, S.: Component software: beyond object-oriented
programming. Addison-Wesley Professional (2002)
25. Zhou, C., Ramacciotti, S.: Common criteria: Its limitations and advice on improve-
ment. Information Systems Security Association ISSA Journal, 24–28 (2011)
Using Future Internet Infrastructure
and Smartphones for Mobility Trace Acquisition
and Social Interactions Monitoring
Athanasios Antoniou, Evangelos Theodoridis, Ioannis Chatzigiannakis,

and Georgios Mylonas
Computer Technology Institute and Press

Rio, Patras, Greece, 26504
{antoniou,theodori,ichatz,mylonasg}@cti.gr
Abstract. Recent activity in the field of Internet-of-Things experimen-

tation has focused on the federation of discrete testbeds, thus placing
less effort in the integration of other related technologies, such as smart-
phones; also, while it is gradually moving to more application-oriented
paths, such as urban settings, it has not dealt in large with applications
having social networking features. We argue here that current IoT in-
frastructure, testbeds and related software technologies should be used in
such a context, capturing real-world human mobility and social network-
ing interactions, for use in evaluating and fine-tuning realistic mobility
models and designing human-centric applications. We discuss a system
for producing traces for a new generation of human-centric applications,
utilizing technologies such as Bluetooth and focusing on human inter-
actions. We describe the architecture for this system and the respective
implementation details presenting two distinct deployments; one in an
office environment and another in an exhibition/conference event1 with
103 active participants combined, thus covering two popular scenarios
for human centric applications. Our system provides online, almost real-
time, feedback and statistics and its implementation allows for rapid and
robust deployment, utilizing mainstream technologies and components.
1 Introduction
Experimentation in the field of Internet-of-Things has currently grown to encom-
pass enormous infrastructure sizes, heterogeneous pools of resources, as well as a
large breadth of application scenarios. Research projects such as WISEBED [1]
and SmartSantander [2] serve as examples of the aforementioned advancements,
depicting the use of federated testbeds of large scale, diverse application scenar-
ios and enormous scale deployment and operation in urban settings. However,
certain aspects of current technology and application trends have not been ef-
fectively dealt with; namely, the use of smartphones in combination with IoT
infrastructure and, on the application side, human mobility and social network-
ing related themes. Instead, the currently utilised application scenarios revolve
1
FET’11, The European Future Technologies Conference and Exhibition.

118 A. Antoniou et al.
less around human activity and more around monitoring environmental param-
eters; opening up to additional possibilities with regard to IoT experimentation
should provide further insight to the Future Internet.
On the one hand, smartphones are increasingly getting closer to the Internet-
of-Things, encompassing at the same time an impressive range of integrated sen-
sors: accelerometers, cameras, gyroscopes, microphones, thermistors, proximity
sensors, etc., while also adopting novel technologies like Near Field Communi-
cation (NFC). Also, the latest smartphone operating systems also offer enough
flexibility for adding external sensing units directly or communicating wirelessly
with them. Furthermore, additional functionality and more potent hardware is
bridging the gap in capabilities with traditional computing systems.
On the other hand, inferring social and contextual interactions has direct and
important applications in our daily lives, uncovering fundamental patterns in
social dynamics and coordinated human activity. Deriving accurate models for
human activity [4] is of great importance to both social sciences and computer
scientists dealing with studies/simulations of mobile users; real-world data can
aid tremendously in this direction, since they can provide a realistic means to
(re)produce, fine-tune and empirically validate models that attempt to capture
the characteristics of human mobility in a variety of environments and situa-
tions/occasions. Similarly, recording the daily activity of elders at home using
sensors can produce patterns that may help in providing a better quality of life
for them. RFID deployments inside a university or enterprise building can reveal
communication patterns among students and faculty over time, helping in un-
derstanding (in)efficiencies in that respect. Smartphones’ proliferation can also
aid in delivering similar results [3]. Finally, an interesting issue is to capture, in
a qualitative and quantitative manner, the characteristics of meetings, confer-
ences and gatherings where a large amount of people from different backgrounds,
disciplines and interests congregate and cooperate.
Therefore, we believe that there is currently a need to add the following per-
spectives to the Future Internet research agenda and develop:
– architectures and systems for combined experimentation using smartphones
and Internet-of-Things devices,
– techniques for sensor-based behaviour profiling and models of behaviour,
– tools that exploit cross-correlations of behavioural profiles of an individual
user and across user groups in order to gain new insights and utilise them
in selected services and applications of high socio-economic value.
We envisage a domain of Future Internet applications that become possible uti-
lizing semantically rich information derived from real-world mobility and pres-
ence traces. Such applications can have as their main focus to perform statistical
analysis and provide reports on collected trace data inferring possible interac-
tions among the monitored population. Other ones can analyse the trace data
and publish results while the data are still being gathered. Additional applica-
tions could use the trace data to predict the future behavior of the observed
population, or even extend the results to larger populations. We also consider
applications that combine a subset or all of the above functionalities, providing
Using Future Internet Infrastructure and Smartphones 119
reports on collected data, generating real-time content in parallel with the trace
gathering process and predicting the behavior of the monitored population.
Moreover, people in cities work in enterprises, offices, etc., spending consid-
erable time inside such environments. Capturing the collaborative, social and
physical behaviour in an organizational context is a critical foundation for man-
aging information, people, and ICT. E.g., customers can be segmented on the
basis of common or similar patterns along multiple behavioural feature dimen-
sions such as frequency of face-to-face contacts, commonality of location and
similarities in movement patterns, as well as commonness in network and ser-
vice use. According to the information richness theory, face-to-face interactions
are the richest and the most effective medium during daily interactions. These
can provide clues of higher quality of social relationships than co-presence indi-
cations, leading to better predictive models about user behaviour. These models
can be utilised for improving current mobility models of mobile subscribers or
consumer models in mobile commerce environments. Furthermore, personalised
content streaming, satisfying customer needs and further pushing business ac-
tivity could be possible by utilising such social networking knowledge, location
awareness and recorded data. Additional examples of such applications are a
smart mall application that can adaptively push product advertisements and
personalised bargain offers to potential customers that move within its premises
and a smart conference scenario, whereby interaction statistics and a presence
heatmap are generated periodically and reported.
Related to such concepts, we discuss here a system for monitoring large groups
of users using a combination of static and mobile IoT infrastructure, target-
ing multiple application domains, which become possible or are considerably
enhanced by analyzing the inferred interactions in space-time-social character-
istics dimensions and furthermore exploiting the prediction of future behavior
and contacts for individuals or groups of people with common social attributes.
Moreover, one should consider our approach in light of the Future Internet vi-
sion and current trends such as crowdsourcing and social computing; we expect
such enablers to unlock the potential of the Internet-of-Things, since computing
is rapidly becoming an integral part of our society. Future systems will orches-
trate myriads of nodes, web services, business processes, people and institutions;
inferring social interactions is needed to support such a Future Internet vision.
We applied our system in 2 scenarios, an office building and a large confer-
ence setting (FET’11) and the results show definite potential in our approach.
We present our architecture and current implementation, along with technical
issues related to our design choices. Along with the monitoring and archiving
functionality of the system, we additionally offer on-line statistics for various fea-
tures. The proposed solution, considers detection of human interaction and pref-
erences by exploiting Internet of Things infrastructures and novel middle layer
mechanisms. We believe that building applications, by adopting the proposed
methodology, can leverage innovation capabilities to a wide range of application
domains like Smart Cities and Smart Organizations.
1.1 Related Work
An early approach in monitoring the mobility of people or classes of people,

the congregation and the interactions among them, was discussed in [5] while in
[6] wearable Bluetooth-enabled devices were used. In [7] trials were conducted
during CoNEXT’06 and FIRE’07. Subsequent works like [9] focus on the uti-
lization of mobile phones and trials in urban settings. In [10] the authors study
data transfer opportunities between wireless devices carried by humans and they
observe that the distribution of the inter-contact time (the time gap separating
two contacts between the same pair of devices) may be well approximated by a
power law over the range.
The recent availability of large-scale datasets, such as mobile-phone records
and GPS data, has offered researchers detailed patterns of human behavior. [12]
studied human movements using a large quantity of bills, while [13] used mobile
phone data from 100K individuals. It was shown that each person is charac-
terized by a time-independent travel distance and a large probability to revisit
previously-traversed locations. In [14] the authors introduce two principles that
govern human trajectories, exploration and preferential return, which are both
unique to human mobility. In [15] individuals’ daily routines are proved highly
predictable, using principal component analysis. Human contact prediction has
also attracted much interest; data mining in social network data (human con-
tact graph) is quite challenging due to the great imbalance in the number of
positive and negative cases in training datasets. Most research efforts propose
various proximity measures on network topology, to be used as predictors for
new contact links [16]. Furthermore, [17] explores the impact of human mobility,
as an intrinsic property of human behavior, on contact link prediction. Datasets
consisting of parallel geographical, network and contact information are scarce,
even today. In [17,18] it is observed that the probability of forming a social tie
decays with distance as a power law. Based on that, in [19] the authors propose
a method that predicts the location of an individual.
In [20] it is stated that there is need for a precise specification of interaction
behaviour in organizations, as information systems require a precise specification
for handling all possible situations. They claim that such interaction behaviour
is described in business process choreographies, a collection of business steps
taking place in a prescribed manner and leading to business objectives. They
conclude that using ICT is crucial for designing and developing tools that will
allow managers to analyze, synthesize and evaluate ways of managing people,
information and technology in public and private sector organizations. Utiliz-
ing IoT and pervasive systems in such context further expands the possibilities
for real-world and real-time applications that can increase the knowledge about
an organization’s process intelligence and thus the efficacy of decision making.
In [21] the analysis of behavioural signals obtained by wearable badges at the
workplace such as face-to-face interactions and modeling the relationship be-
tween organizational dynamics and organizational performance based on that,
is shown to be an effective management tool that can radically improve the
overall operation of the organization.
2 A System for Trace Acquisition - Architecture and

Implementation
Our architecture for collecting traces of presence in a Smart City environment
is partitioned in 3 tiers. The lower tier contains the fixed location base station
trackers and the mobile personal devices (i.e., mobile phones) carried by the
monitored population. The mobile personal devices are further divided in simple
devices which can be detected by the base station trackers, and mobile trackers
which are capable of detecting other mobile devices and nearby base stations.
The base station trackers are placed at fixed locations throughout the mon-
itored area providing the coverage required. These trackers are interconnected
using a reliable and sufficiently high-throughput technology (e.g., 100Mbps Eth-
ernet, 802.11g). We currently use the Bluetooth in our enabling devices - it is
a ubiquitous technology with which end-users feel comfortable, while IoT nodes
and smartphones also usually support it. The scan range of a tracker is typically
10-20 meters, but the system does not impose a specific constraint and can sup-
port trackers with varying scan ranges. In most cases, the trackers are placed
in proximity of each other so that their scan ranges overlap. In this way, we are
able to infer presence of a device at intermediate locations using the received
signal strength within a short-time window. Each base station tracker maintains
a local log of detected traces in addition to forwarding them towards the local
(on site) database. The mobile trackers are utilised to complement the static in-
frastructure and collect additional traces of mobile phones, even when those are
located outside the range of the static base stations. They periodically attempt
to transmit their buffered trace data via a WiFi connection to the Application
Server, which in turn relays this data to the remote DB Server.
In the second tier, the collected traces from each static tracker are stored in a
local database - essentially records of device traces with a corresponding inquir-
ing tracker ID, a timestamp and a RSSI value. These data are also forwarded
to the remote database and analysis server, where they are used to produce
meaningful results. The Remote Database and Analysis Server is typically ac-
cessible over the Internet via a secure connection channel. In the preprocessing
stage the trace data are filtered to remove duplicate and invalid entries as well
as entries from devices not participating in the monitoring system. Furthermore,
for each trace, a specific location is assigned to the mobile device and hence to
the person carrying it, by considering the RSSI of the device as measured from
involved base stations in a short-time interval around the trace timestamp. The
remote DB adopts a more advanced schema that allows taking into considera-
tion a time-schedule of events in different monitored locations, the participants’
interests and personal attributes (e.g. age range, scientific background). During
the analysis phase, possible interactions among the population are inferred and
correlated with their self-reported attributes and scheduled events.
The third tier, is essentially the application layer of the architecture. A web
site provides information about the related deployed monitoring application, a
description of the system technology, instructions for participation and links to
interesting results from the traces analysis. The system will only process traces
that correspond to the presence of people who agreed to participate and carry
mobile detectable devices, submitting a registration form. In addition to the
participation consent, the registration form may request optional information
regarding personal attributes of a participant, that will be used to infer behav-
ioral patterns for groups of people that share common attributes. The application
layer also includes an automated mechanism that posts links to interesting re-
sults with a short description on a Twitter account, which end-users can follow
in order to receive updates about the dynamics of the participants’ interactions.
Our system architecture was designed with an emphasis on scalability, ease of
deployment and simplicity in participation requirements, as well as the ability
for people to register online, even after the monitoring deployment has launched.
Such flexibility is usually absent in other related systems, both in terms of adding
users online, as well as modifying the supporting infrastructure and maintaining
overall system stability. The distributed nature of our system results also to an
easier and faster installation phase.
Another basic consideration was respecting the privacy and self-reported data
of the participants, and the deletion of “external” traces belonging to devices
not registered for the particular deployments. Privacy concerns of the partici-
pants were answered by the anonymisation of collected data. Since privacy is-
sues should not be perceived by the participants themselves as an afterthought,
all were informed prior and during the experiments regarding the data col-
lection aspects of their participation, the future availability of the produced
anonymised data sets and our conformance to the related legislation (EU direc-
tive 95/46/EC). At the same time, users had control over the software compo-
nents running on their smartphones and could opt for turning them off anytime.
By utilizing Bluetooth networking, which is supported by the vast majority
of the mobile phones that are in use today, certain advantages were evident:
participants are only required to carry with them a personal device, the collected
trace data can be delivered in real-time, while also the infrastructure cost is cheap
to purchase and maintain. Moreover, Bluetooth allows for greater localisation
accuracy compared to WiFi, due to its more limited range. It is also easier and
safer to setup and operate, due to the inherent features in Bluetooth’s design.
2.1 Implementation Details

Mobile Trackers: The mobile trackers are used to complement the static in-
frastructure and collect additional traces of mobile phones, even when those are
located outside the range of the static base stations. The mobile trackers in
our implementation are Nokia smartphones with Bluetooth and WiFi support
and Android based phones. The mobile application has a simple GUI and of-
fers the option of running hidden as a background application. A mobile tracker
performs a periodic inquiry scan for discoverable Bluetooth devices, i.e., users’
mobile phones and static base stations. The list of detected devices is stored on
a local limited-length buffer of the “active sessions”. Each entry contains two
timestamps, for the first and last time the device was encountered. If a previously
detected device is not seen in a new scan, then its entry is moved in another
local buffer ( “completed sessions”). Occasionally, the mobile tracker attempts

to forward stored sessions to the application server that handles these traces
(through WiFi). During such an opportunity, the completed session traces are
transmitted first, followed by the active sessions traces. By running the mobile
tracker application, a mobile phone is able to detect the base station trackers
and can therefore be associated with them without having to operate in discov-
erable mode. The application server inserts the data from the mobile trackers
into a separate table in the remote database.
Basestation Trackers: The base station trackers are hosted on mini PCs
or laptops that have Bluetooth dongles attached and are placed at specific fixed
locations, providing full coverage for a monitored area. In most cases, the track-
ers are intentionally placed in proximity of each other so that their scan ranges
overlap. This way we are able to infer presence of a device at intermediate lo-
cations by evaluating the received signal strength from the same device at each
tracker within a short-time window. All trackers are time-synchronized with the
local DB server, their Bluetooth interfaces are set in discoverable mode and pe-
riodically perform a Bluetooth scan inquiry of a predefined duration. For each
detected device a trace entry is created, including a timestamp and a corre-
sponding RSSI value. The traces gathered after an inquiry scan are transmitted
towards the local database after the end of the inquiry phase.
Local Database, Remote Database and Analysis Service: We used
a MySQL instance on a laptop, while mobile and base station trackers record
the users’ presence directly in this database instance. Every 5 minutes this local
instance pushes the updates, utilizing a CRON daemon and SSH connection to a
mirrored MySQL instance hosted at the remote database machine. This remote
database server was hosted at the headquarters of CTI and this two step schema
was used due to unstable Internet connection and limited processing resources
on the devices utilized on site. Our Remote Database consisted of a MS SQL
Server 2008. Services deployed and used were MSSQL Server RDBMS, MSSQL
Integration Services and MySQL. All functionality and instrumentation at the
centralized server was implemented by a set of tasks in the MSSQL Integration
Services. Whenever an update took place in the local MySQL instance, all trace
records were retrieved and forwarded for processing. Initially, the MAC addresses
in the trace records were removed and replaced by a user ID that was correlated
with the social attributes of the users. Thus, the subsequent Aggregation and
Analysis phases were not aware of the user MAC.
A location ID is then assigned to each trace record in order to verify co-
presence of users and attendance to events. Using a 1-minute buffer, we lookup
the reachable base station trackers for each user and respective RSSI values
and form an “observed vector” for the user during that interval. From a set
of possible vectors, mapping base stations to sublocations and indicating the
trackers’ capability of detecting presence for devices, we pick the one more similar
to the observed vector. This set of vectors is recorded in a training phase. At
the end of this step all trace records were quantized at a 1-minute time interval
grain, characterized by location. This set of trace records was used as a fact table
in order to be analyzed in a MOLAP Cube in MSSQL Integration Services. Fact

records were analyzed by date dimension in a Day, Hour, Minute hierarchy, by
“Persons” dimension with social characteristics attributes (age range, profession,
etc.), by “Place” and “Event” dimensions. The Event dimension is a function
dimension on the “Place” and “Date” dimensions.
3 Deployment and Results Discussion
We deployed our system twice: a) inside our institute building, with 23 Bluetooth-
capable base stations, distributed over 3 building levels, monitoring for 27 hours
(9am-6pm, 3 days), b) at a large-scale conference event, with 36 base stations
(12 mobile), for 27 hours. A total of 103 participants in both events carried with
them their mobile phones, with Bluetooth switched on, set to be discoverable.
We describe here the main characteristics of the discussed deployments.
CTI Deployment: In essence, a building-scale IoT infrastructure was used
to monitor interactions between co-workers and/or different enterprise depart-
ments, in order to infer both online and over time intraconnections and in-
terconnections within such entities. This kind of knowledge could give further
insight for optimizing business processes, re-organizing hierarchical structures or
re-establishing connections through e.g., reimplementing certain standard proce-
dures or changing the actual physical locations of specific people or departments.
CTI’s staff consists of a number of research teams and administrative / support
staff, with each one housed in discrete parts of the CTI building. Moreover, CTI
is situated in a 5-floor building, with the thick walls and steel doors of each floor
sector providing isolation in terms of wireless communication between adjacent
parts. This provided an advantage in determining the position of participants
inside the building more accurately. The setup of the system inside 23 different
building rooms overall required 4 hours of work from 3 members of our team.
Bluetooth-enabled gateways were used in all rooms, being powered on for the
whole duration of the experiment, monitoring all Bluetooth networking activ-
ity and reporting to the system, as defined in Section 2. The duration of the
experiment prohibited the use of battery-powered gateways, since we wanted
the infrastructure to operate largely unattended. The layout of the building
also contributed in confining the activity of people interested only in commu-
nicating within their own group, allowing the activity of persons behaving as
“hubs” between different groups to be more visible. It is interesting to note
that we monitored physical presence, and thus interaction in the physical space.
As discussed in the next section, it reflects the structure of the institute quite
accurately.
FET’11 Deployment: In the second set (conference) of experiments, a number
of weeks after our initial deployment, we tested our system in a less controlled en-
vironment. The performance fine-tuning after the first set of experiments allowed
us to scale the system even further. Since this was a larger scale deployment and
was done in harsher conditions, we used a larger team of people to setup and
operate the infrastructure, i.e., 5. The setup was completed within 4 hours on
the day before the opening of the conference (FET’11) to the public. In contrast
with the CTI deployment, the networking isolation offered by walls and doors
was not available in this case, making it more difficult to determine the location
of participants. Furthermore, we implemented additional components for provid-
ing results and statistics, e.g., posting latest information about booth popularity
on Twitter and other social networks. Conference participants showed interest
towards the statistics, even though they were produced with some minutes of
delay. The set of statistics produced included information such as the top 10 pop-
ular booths, booths where visitors spend the most time, among others. Apart
from visitors, exhibitors also showed interest in statistics about their booth and
indicators regarding how their exhibits faired against others.
3.1 Results Discussion

One of our basic findings during deployment and operation of our system is that
it is possible to acquire and process human mobility data, extract human interac-
tions and analyse them in almost real time manner, combining widespread tech-
nologies and relatively simple and low-cost subsystems. The flow of human trace
data from lower infrastructure components is channeled to the web, enriched
with the inferred human interaction possibilities and self-reported personal in-
formation (e.g., age, profession, interests). By exposing this rich flow of data,
new opportunities arise to build interesting applications on top of it (like real
world recommendations, searching and discovering people of different knowledge
backgrounds and social profiles etc.) or design interconnecting testbeds exchang-
ing complex analyzed information in addition to plain data messages. In such
cases, the system design should perceive the information analysis of human in-
teractions as part of a communication protocol running concurrently on top of
heterogeneous resources. In both deployments, the assumption that users carry
their smartphones constantly with them was largely confirmed, while commu-
nication with an IoT infrastructure in a pervasive manner helped to ensure the
correct operation of the system with minimal user time consumed.
During the first set of experiments in the CTI building, a series of communi-
cation patterns among participants emerged. Fig. 1 (a) reveals a dense network,
depicting the gradual cooperation of users. The diameter of the network, was
found to be 3, meaning that any two members of the personnel can either com-
municate directly (one hop), or through at most two intermediate people. These
reflect a hierarchical administration structure and strong interaction. Fig. 1 (b)
depicting the contact network parameters shows that the network centralization
and network heterogeneity is average while the clustering coefficient is quite
large, indicating that the contact graph tends to form a clique. Fig. 2(b) de-
picts interactions of the participants within their own and among other groups
(research units) as well. We can observe clearly that some groups have strong
intraconnections and strong interconnections with some of the other groups.
Such information can be used e.g., in an enterprise to detect inefficiencies in its
management structure, or evaluate potential solutions immediately and express
(a) (b)
Fig. 1. (a) Rising number of nodes and average degree reflects the users’ gradual in-
volvement. (b) The general features characterizing the graph can be captured in rela-
tively little time in both scenaria.
explicitly mobility habits among sections of the institute. Finally, in Fig.2 (a)
we also show the intensity of interaction among research units.
In the second set of experiments (FET conference setting), we observed slightly
different behaviors. While Fig. 1 (a) reveals for this case a similarly dense net-
work and the diameter of the network is again found to be 3, Fig. 1 (b) depicts a
greater tendency to form a clique as the clustering coefficient and network den-
sity is large (larger than the CTI deployment). Network centralization is quite
low (lower than CTI) while Network Heterogeneity is average again. Fig. 3(b)
depicts interactions of the participants among groups of different scientific back-
ground. In Fig. 3(a) the distinct number of users who attended each of the 10
most popular booths for each day of FET’11 is presented, while Fig. 4 depicts
the average time spent by each Scientific Background group in each booth. Such
information was delivered almost online, i.e., with a latency of about 5 minutes,
and can be utilised in accessing overall tendencies in such an event and deliv-
ering useful statistics to both participants and organisers. Overall, the statistics
delivered could reveal “hidden” trends and synergies between different scientific
fields, which could otherwise be difficult to recognise.
4 Conclusions - Future Work

We believe that recent progress in human mobility modeling and the rise of
applications with social networking characteristics should be encompassed in
current IoT experimentation activities. In that respect, the fuse of smartphones
and IoT infrastructure can enable systems such as the one presented here. We
experimented in two discrete scenarios, an office building and a scientific confer-
ence hall and deployed our system to capture human mobility and interactions.
Our future work will focus on extending the current range of supported mobile
platforms and providing a better end-user experience, and also provide traces
on an even larger scale, such as in a smart IoT city setting.
Educational Technology
Project Management
Division of Technical Support
Research Unit 1
e-Government
Administrative and Financial Dpt
Computing and Networking Services
(a) (b)
Fig. 2. Interactions among (a) the various groups in CTI - the results reflect largely the
hierarchical structure and actual cooperation patterns among groups, (b) interaction
matrices reveal strong cooperation among participants’ own groups at all times, with
varying degree during different times of the day
Environmental science
Chemistry Other
Mathematics
Biology
null
Computer sciences
Engineering
Economics Physics
Psychology
Management
(a) (b)
Fig. 3. (a) Number of distinct users per booth per day, (b) various groups of partici-
pants in FET’11 with different scientific background
Fig. 4. Average Time (min) each Scientific Background group spent in each booth
Acknowledgments. This work has been partially supported by the European

Union under contract number ICT-258885 (SPITFIRE).
credited.
References
1. Coulson, G., et al.: Flexible experimentation in wireless sensor networks. Commu-
nications of the ACM (CACM) 55(1), 82–90 (2012)
2. SmartSantader project, http://smartsantander.eu
3. Miluzzo, E., et al.: Darwin Phones: the Evolution of Sensing and Inference on
Mobile Phones. In: MobiSys 2010, pp. 5–20 (2010)
4. Eagle, N., Pentland, A.: Reality Mining: Sensing Complex Social Systems. In: Per-
sonal and Ubiquitous Computing, pp. 255–268 (May 2006)
5. Borovoy, R., et al.: Meme tags and community mirrors: moving from conferences
to collaboration. In: CSCW 1998, pp. 159–168 (1998)
6. Hui, P., et al.: Pocket switched networks and human mobility in conference envi-
ronments. In: WDTN 2005, pp. 244–251 (2005)
7. Nordstrom, E., Diot, C., Gass, R., Gunningberg, P.: Experiences from measuring
human mobility using Bluetooth inquiring devices. In: MobiEval 2007, pp. 15–20
(2007)
8. Nicolai, T., Yoneki, E., Behrens, N., Kenn, H.: Exploring Social Context with the
Wireless Rope. In: Meersman, R., Tari, Z., Herrero, P. (eds.) OTM 2006 Workshops,
part I. LNCS, vol. 4277, pp. 874–883. Springer, Heidelberg (2006)
9. Natarajan, A., Motani, M., Srinivasan, V.: Understanding Urban Interactions from
Bluetooth Phone Contact Traces. In: Uhlig, S., Papagiannaki, K., Bonaventure, O.
(eds.) PAM 2007. LNCS, vol. 4427, pp. 115–124. Springer, Heidelberg (2007)
10. Chaintreau, A., Hui, P., Crowcroft, J., Diot, C., Gass, R., Scott, J.: Impact of
Human Mobility on Opportunistic Forwarding Algorithms. IEEE Transactions on
Mobile Computing 6(6), 606–620 (2007)
11. SocioPatterns, http://www.sociopatterns.org
12. Brockmann, D., Hufnagel, L., Geisel, T.: The scaling laws of human travel. Na-
ture 439(7075), 462–465 (2006)
13. Gonzalez, M., Hidalgo, C., Barabasi, A.-L.: Understanding individual human mo-
bility patterns. Nature 453(7196), 779–782 (2008)
14. Song, C., Koren, T., Wang, P., Barabasi, A.-L.: Modelling the scaling properties
of human mobility. Nature Physics (2010)
15. Eagle, N., Pentland, A.: Eigenbehaviors: Identifying structure in routine. Behav-
ioral Ecology and Sociobiology 63(7), 1057–1066 (2009)
16. Liben-Nowell, D., Kleinberg, J.M.: The link prediction problem for social networks.
In: CIKM 2003, pp. 556–559 (2003)
17. Wang, D., Pedreschi, D., Song, C., Giannotti, F., Barabasi, A.-L.: Human Mobility.
In: Social Ties, and Link Prediction KDD 2011 (2011)
18. Lambiotte, R., et al.: Geographical dispersal of mobile communication networks.

Physica A: Statistical Mechanics and its Applications 387(21), 17 (2008)
19. Backstrom, L., Sun, E., Marlow, C.: Find Me If You Can: Improving Geographical
Prediction with Social and Spatial Proximity. North, pp. 61–70. ACM (2010)
20. Decker, G., Weske, M.: Interaction-centric modeling of process choreographies. Inf.
Syst. 36, 292–312 (2011)
21. Olguin, D.O., et al.: Sensible Organizations: Technology and Methodology for Au-
tomatically Measuring Organizational Behavior. IEEE Transactions on Systems,
Man, and Cybernetics 39 (February 2009)
I-SEARCH: A Unified Framework
for Multimodal Search and Retrieval
Apostolos Axenopoulos1, Petros Daras1, Sotiris Malassiotis1, Vincenzo Croce2,

Marilena Lazzaro2, Jonas Etzold3, Paul Grimm3, Alberto Massari4, Antonio Camurri4,
Thomas Steiner5, and Dimitrios Tzovaras1
1
Centre for Research and Technology Hellas (CERTH/ITI),
6th Km Charilaou-Thermi, 57001 Thessaloniki, Greece
{axenop,daras,malasiot,tzovaras}@iti.gr
2
Engineering Ingegneria Informatica S.p.a.,
Viale Regione Siciliana Nord Ovest, 7275, 90146, Palermo, Italy
{vincenzo.croce,Marilena.Lazzaro}@eng.it
3
Hochschule Fulda, Marquardstrasse 35, 36039 Fulda, Germany
{paul.grimm,jonas.etzold}@informatik.hs-fulda.de
4
Casa Paganini – InfoMus Research Centre, University of Genoa,
Piazza S.Maria in Passione 34, Genoa, Italy
{alby,toni}@infomus.org
5
Google Germany GmbH, ABC-Str. 19, 20354 Hamburg, Germany
[email protected]
Abstract. In this article, a unified framework for multimodal search and

retrieval is introduced. The framework is an outcome of the research that took
place within the I-SEARCH European Project. The proposed system covers all
aspects of a search and retrieval process, namely low-level descriptor
extraction, indexing, query formulation, retrieval and visualisation of the search
results. All I-SEARCH components advance the state of the art in the
corresponding scientific fields. The I-SEARCH multimodal search engine is
dynamically adapted to end-user’s devices, which can vary from a simple
mobile phone to a high-performance PC.
Keywords: multimodal search, multimodal interfaces, adaptive presentation.
1 Introduction
Current Internet (CI) was developed 30 years ago for serving research demands (host-
to-host communications). However, it is obvious that it cannot be used today with the
same efficiency, since new demanding applications rise. The number of Internet users
as well as the available multimedia content of any type increase exponentially.
Moreover, the increase of user-generated multimedia content and the number of
mobile users will raise new challenges. Towards this direction, the Future Internet
(FI) aims to overcome current limitations and address emerging trends including:
network architectures, content and service mobility, diffusion of heterogeneous nodes
and devices, mass digitisation, new forms of user-generated (multimodal) content
I-SEARCH: A Unified Framework for Multimodal Search and Retrieval 131
provisioning, emergence of software as a service and interaction with improved

security, trustworthiness and privacy [1].
With respect to content characteristics, the content supported by FI could be:
intelligent, i.e. able to be adapted to the user preferences, devices and access
networks; 3D and haptic, including also visual and sound features, as well as
physiological or emotional user’s state; interactive, allowing user to interact with the
media objects; cross-modal and multimodal, thus, providing intuitive links among
future media and enabling search and retrieval from one modality to another; and
collaboratively edited/filtered, allowing editing, filtering and manipulation of content
in a collaborative way. FI is expected to address several limitations of CI, with respect
to content, such as disembodied and non-multimodal access to content. The lack of
embodiment in CI could be faced by enhanced support of multimodality, including
sound, haptics, visual, gestural, physiological, toward a deeper exploitation and
integration of communication and interaction through the physical, non-verbal, full-
body channels [1].
In this sense, the EU-funded project I-SEARCH aims to create a unified
framework for multimodal search and retrieval, which is fully inline with the vision
and objectives of FI. The search engine proposed by I-SEARCH enables retrieval of
several types of media (3D objects, 2D images, sound, video and text) using as query
any of the above types or their combinations. The framework provides novel
multimodal interaction mechanisms to enable easy retrieval and access by users to
multimedia content as well as to capture the emotional expressive and social
information conveyed by both individual and groups of expert and non-expert users.
Moreover, it provides novel data representations and transformations in order to
support conversion of all types of conflicting and dynamic data in ways that support
visualization and analysis. Finally, it provides device adaptation capabilities,
addressing several types of end-user devices, such as PCs, mobile phones, PDAs and
smart phones. In this paper, the overall architecture and main functionalities of the I-
SEARCH framework are presented.
1.1 Related Work
While the problem of retrieving one single modality at a time, such as 3D objects,
images, video or audio has been extensively covered, retrieval of multiple modalities
simultaneously (multimodal retrieval) has yet to yield significant results. In [10], the
intra- and inter-media correlations of text, image and audio modalities are
investigated in order to produce a Multi-modality Laplacian Eigenmaps Semantic
Subspace (MLESS). In [11], a structure called Multimedia Document (MMD) is
introduced to define a set of multimedia objects (images, audio and text) that carry the
same semantics. After creating a Multimedia Correlation Space (MMCS), a ranking
algorithm is applied, which uses a local linear regression model for each data point
and it globally aligns all of them through a unified objective function. Within I-
SEARCH, an approach for multimodal retrieval has been introduced. It is based on
Laplacian Eigenmaps [12], while it has been further enhanced with large-scale
indexing [13] and relevance feedback [14].
132 A. Axenopoulos et al.
The integration of non-verbal expressive, emotional and social dimensions in

multimodal queries enables novel ways users can access content. In [16] a novel
paradigm for modelling and analyzing non-verbal full-body affective gestures is
proposed. An approach to model and analyse full-body non-verbal social signals
(entrainment, leadership) is presented in [17].
Multimodal search engines are still very experimental at the time of writing. For
our work on I-SEARCH, we looked for common patterns in search-related actions.
Across the Web, the pattern that is used for almost all search related actions is the text
field. From big Web search engines such as Google, Yahoo, or Bing, to intranet
search engines, this pattern stays the same. However, I-SEARCH cannot directly
benefit from this broadly accepted pattern, as a multimodal search engine must
support a large number of query types simultaneously: audio, video, 3D, image, etc.
Some current search engines, even if they do not have the need for true multimodal
querying, still do have the need to accept input that is not plain text. As a first
example, we consider TinEye1. TinEye is a Web-based search engine that allows for
query by image content (QBIC) in order to retrieve similar or related images. The
interface allows for direct file upload, however, the requirements for a multimodal
search engine like I-SEARCH are more complex. As a second example, we examine
MMRetrieval [6]. It brings image and text search together to compose a multimodal
query. MMRetrieval is a good showcase for the problem of designing a UI with many
user-configurable options as well as multimodal aspects. For a user which is not
involved within the field of information retrieval, the UI seems not necessarily clarify
the meaning of all inputs in detail, especially when field-specific terms are used.
Finally, we have a look at Google Search by image2, a feature introduced in 2011 with
the same UI requirements as MMRetrieval: combining text and image input. With the
Search by image interface, Google keeps the text box pattern, while preventing any
extra visual noise. The interface is exposed to users via a contextual menu when the
camera icon is clicked.
Independently of the techniques used for querying and retrieval of multimedia
databases, presentation of the results follows similar patterns as with text search.
Major search engines such as Google Images and Bing Images present results as a
rectangular grid or matrix of thumbnails that are ordered from left to right and top to
bottom based on their ranking score. Google Videos and Youtube present results as a
linear list of video surrogates containing a representative video shot plus
accompanying text summary and metadata. Also, numerous interfaces have been
developed for image browsing of personal collections. For example, in the PhotoMesa
image browser [7], images in a directory are arranged in space filling boxes using a
quantum Treemap algorithm. Clustering images by time is a popular way for
organisation of personal collections [8]. In PhotoTOC [9] content based clustering is
applied after time-based clustering for clusters that contain many images. Clustering
based on faces was recently introduced in applications such as Google Picasa, Apple
iPhoto and Flickr.
1
http://www.tineye.com/
2
http://www.google.com/insidesearch/searchbyimage.html
2 Overview
In multimodal search and retrieval problems, it is much more convenient to enclose

multiple media types, which share the same semantics, into a media container, and
label the entire container with the semantic concept, instead of labelling each media
instance separately. Following this approach, in I-SEARCH, the concept of Content
Object (CO) has been introduced to describe such rich media containers. A CO can
span from very simple media items (e.g. a single image or an audio file) to highly
complex multimedia collections (e.g. a 3D object accompanied with multiple 2D
images and audio files). Moreover, a CO may include additional metadata related to
the media, such as textual information, classification information, real-world data
(location or time-based), etc. When a user refers to a CO, s/he directly refers to all of
its constituting parts. A detailed description of CO is available at [1].
Fig. 1. Block diagram of the I-SEARCH framework
A block diagram of the I-SEARCH framework is given in Fig. 1. During the

offline phase, the COs of the I-SEARCH dataset are inserted to the Content Analytics
Controller (CAC). CAC is responsible for extracting low-level descriptors for each of
the CO’s constituting modalities. The output is a set of low-level descriptors, which
are stored using a novel description format called Rich Unified Content Description
(RUCoD). The RUCoD format is also analysed in [1]. RUCoD descriptors are
processed using a novel manifold learning framework, producing a set of multimodal
descriptors, which are efficiently indexed to facilitate faster retrieval.
During the online phase, the user initiates a multimodal search session by adding
one ore more modalities to the appropriate I-SEARCH interface. The interface
supports text, image, video, audio and 3D queries, as well as emotional (captured by
user’s expressions) and real-world (user location, time) input. A query RUCoD is
produced using CAC, which is used to query the multimodal index. The retrieved
ranked list of COs is optimally presented to the user through the adaptive presentation
component. A description of the I-SEARCH components follows.
3 Content Analytics Process
During the Content Analytics Process, an appropriate analysis is performed in order to

extract descriptors from the CO’s constituting modalities and store them using a
multi-level structure. This structure takes into consideration: i) content-specific low-
level descriptors, which characterise the type of content, ii) real-world descriptors,
which associate the content with information extracted from sensors (i.e. GPS,
temperature, time, weather, etc.), and iii) user-related descriptors, which encapsulate
expressive, social and emotional characteristics to the semantics of these items.
The Content Analytics Controller (CAC) is the process orchestrator for low-level
descriptor (LLD) extraction. As a result, LLDs are extracted for each modality within
the CO and further merged into a RUCoD file. Each RUCoD is the data
representation of a CO and consists of two main parts: Header and Description tags.
The former includes general information edited by Content Providers during the
content injection phase. The latter is representing the CO low-level features for each
multimedia, real-world and user-related information. Moreover it contains the
artefacts (i.e. thumbnails, key-frames, etc.) that are produced as intermediate results
of low-level feature extraction phases.
Specific RUCoD Tools have been developed for content injection and RUCoD
header production, which is the preliminary part of the Content Analytics process:
The RUCoD Authoring Tool (RAT) supports CO creation from existing media
collections. It takes as input all different types of media items, real-world information
and user-related information (emotional/expressive characteristics); as results a rich
media representation of the Content Object is produced according to RUCoD format
xml schema.
The Crawler2RUCoD script supports creation of Collection of COs starting from a
corpus of multimedia content. This strategy is an automatic creation of one CO for
each media.
The CoFetch RUCoD Tool performs a semi-automatic creation of COs. It provides
a smart way to create a RUCoD starting from keywords. CoFetch RUCoD Tool
performs search on public media sources (Text, 3D, Image, Audio and Video) and
creates corresponding COs.
The core of CAC process comprises a first phase of identification of multimedia
content types followed by triggering of the corresponding LLD extractors. Moreover,
the CAC process is responsible for merging the results of LLD extractors into valid
RUCoD files. As soon as the updated RUCoDs are stored in the platform, the Search
Engine Indexers are notified. Indexers are in charge of retrieving relevant COs during
the online phase.
4 Multimodal Indexing
The low-level descriptors of the COs’ constituting modalities are further processed to
construct a new multimodal feature space. In this new feature space all COs,
irrespective of their constituting modalities, are represented as d-dimensional vectors,
where semantically similar COs lie close to each other with respect to a common
distance metric. The methodology, which is usually followed, is known as manifold

learning, where it is assumed that the multimodal data lie on a non-linear low-
dimensional manifold. The majority of manifold learning approaches is based on the
computation of the k-nearest neighbours among all items of the dataset in order to
create an adjacency matrix. In our case, the items of the dataset are COs. The k-
nearest neighbour computation for a CO is not a trivial process, since it requires
merging descriptors of heterogeneous modalities into one unified distance metric. To
avoid merging of heterogeneous distance metrics, an alternative approach has been
introduced in I-SEARCH [13]. The method is based on Laplacian Eigenmaps (LE)
but, in our case, the creation of the adjacency matrix is modified as follows: when
items i, j are neighbours, the item Wij of the adjacency matrix is assigned the value 1
instead of the actual distance between i and j. A detailed description of the method is
available at [13].
Fig. 2. Block diagram of the I-SEARCH framework
The aforementioned method relies on the calculation of all-to-all distance matrices

among all objects of the dataset. However, when it comes to really large multimedia
datasets, both calculation and storage of all-to-all distance matrices becomes
prohibitive. Consequently, the distance matrix does not provide an efficient solution
in real-life problems. On the other hand, multimedia indexing is a widely used method
to speed up the nearest-neighbour search in large databases. Through indexing, there
is no need to compute one-to-all distances of the query with all database objects. In I-
SEARCH, a new large-scale multimedia indexing approach has been adopted to index
the multimodal descriptors. The main idea of the method is that when two objects are
very similar (close to each other in a metric space) their view of the surrounding
world is similar as well. Thus, instead of using the distance between two objects, their
similarity can be approximated by comparing their ordering of similarity according to
some reference points [15].
5 Multimodal Interfaces
5.1 The I-SEARCH Graphical User Interface and Multimodality
With the I-SEARCH project, we aim at the creation of a multimodal search engine
that allows for both multimodal in- and output. Supported input modalities are audio,
video, rhythm, image, 3D object, sketch, emotion, social signals, geolocation, and text
[5]. Each modality can be combined with all other modalities in an enhanced version
of the search box pattern. The graphical user interface (GUI) of I-SEARCH is not tied
to a specific class of devices, but rather dynamically adapts to the particular device
constraints like varying screen sizes of desktop and mobile devices like cell phones
and tablets. Fig. 3 gives an impression of what this adaptive behaviour looks like in
practice and how multimodal queries are assembled i.e. on a mobile device (Fig. 4).
The I-SEARCH GUI is implemented with the objective of sharing one common code
base for all possible input devices.
Fig. 3. Automatic adaption of I-SEARCH GUI to different devices and screens
It uses a JavaScript-based component called UIIFace [4], which enables the user to
interact with I-SEARCH via a wide range of modern input modalities like touch,
gestures, or speech. Therefore it provides an adaptive algorithm for gesture
recognition along with support for novel input devices like Microsofts Kinect in a
web environment. The GUI also provides a WebSocket-based collaborative search
tool called CoFind [4] that enables users to search collaboratively via a shared results
basket, and to exchange messages throughout the search process. A third component
called pTag [4] produces personalized tag recommendations to create search queries,
filter results and add tags to retrieved result items.
One important goal of I-SEARCH is to hide this complexity from the end-user
through a consistent and context-aware user interface based on standard HTML5,
JavaScript, and CSS, with ideally no additional plug-ins like Flash required. We aim
at sharing one common code base for both device classes, mobile and desktop, with
the user interface getting progressively enhanced [3] the more capable the user's Web
browser and connection speed are. Search engines over the years have coined a
common interaction pattern: the search box. We enhance this interaction pattern by
context-aware modality input toggles that create modality query tokens in the I-
SEARCH search box. Below within Fig. 5, three example modality query tokens for
audio, emotion, and geolocation, can be seen.
Fig. 4. Multimodal query consisting of geolocation, video, emotion and sketch
Fig. 5. Multimodality-enhanced search box pattern with query
5.2 Expressive and Emotional Interfaces

Main innovations proposed in I-SEARCH include (i) the extraction of expressive and
emotional information conveyed by a user to build a query, and (ii) the possibility to
build collective queries, i.e. queries resulting from a social interaction – verbal as well
as non verbal – among a group of users. The I-SEARCH platform includes real-time
algorithms for the analysis of non-verbal emotional behaviour expressed by full-body
gesture, algorithms for the analysis of the social behaviour in a group of users, and
methods to extract data from sensors for accessing real-world information. In the
following we sketch a couple of use cases to explain the need for automated analysis
techniques of non-verbal emotional and social behaviour.
To describe this type of interfaces, we sketch up a couple of use cases, which are
also studied in I-SEARCH: a) Individual multimodal search of music content and b)
Social multimodal search of music content.
According to the first use case, a professional user is looking for music material
that share common features. This research aims at discovering unexpected filiations
and similarities across music artworks. The target group can vary from professional
users/music experts to end-users/music lovers. Multimodal input includes text (words,
phrases, tags, etc.), audio files/clips (query by example), gestures captured via a video
camera or accelerometers embedded in mobile devices and real-world information
(e.g. the GPS position of the user). Typical search and retrieval tasks are the
following: search for a list of audio files having the same rhythm of a pattern specified
by the user (via tapping with a finger on a table/microphone, clapping her hands or
moving her arms in the air) but also sharing the same emotional features (e.g. similar
level of arousal) of the captured user movements or attitude.
The second use case deals with collaborative music retrieval by a group of users.
More specifically, four friends at a party wish to dance together, and to accomplish this
they search some music pieces resonating with their (collective) mood. They do not
know in advance the music pieces they want, and they use the I-SEARCH tool
collaboratively to find their music, and possible associated videos. Multimodal input
includes audio or video clip of a favourite singer (query by example), text-based
keywords, rhythmic queries (using hands, clapping, full-body movement), gestures,
entrainment /synchronization and dominance/leadership among users, measured by on-
body sensors and/or environment video cameras. Typical search and retrieval tasks
include the following: Iterative search for audio files (as well as the video clips or
images that are associated with them) by periodically performing a query for a new
music piece similar to the one currently been played and having a location in the
valence/arousal plane close to the position obtained from the movements of the dancers.
6 Adaptive Presentation
The proposed visualisation framework is based on a hierarchical conceptual
organization of the dataset. According to this conceptual organizations the result of
each query may be diverse enough to be organized in several topics and associated
sub-topics, while each sub-topic (at the bottom of the hierarchy) may be specific
enough to be mapped to a continuous similarity space designating a variability of a
single object along some important dimensions. We argue that such organization is
very suitable for explorative browsing of dataset and is diverse enough to cover a vast
range of data, information needs, and browsing tasks. To achieve the proposed
organization, we automatically augment the results of the multi-modal search engine
with analytics information. In particular, given a mutual similarity matrix among
results documents we perform hierarchical clustering by means of spectral clustering
algorithm. For each resulting group of results we subsequently perform a
dimensionality reduction or transformation algorithm (e.g. minimum spanning trees)
that maps documents on 2D “similarity space”.
We use Treemaps, Hyperbolic Trees and classical tree-like structures
interchangeably to navigate the user to specific groups of results. To avoid cluttered
displays of documents with similar coordinates we employ a fast thumbnail placement

algorithm that is similar to those employed for placing labels on a cartographic map.
For visual multimedia content, such as images, video, 3D objects, an iconic or
pictorial representation of the item, such as an image thumbnail, provides a summary
of the object descriptive enough for the user to make relevance judgments. While
generation of such pictorial representations is straightforward for inherently pictorial
media, it is more difficult with media that are inherently non-visual and/or have a
strong temporal dimension such as audio and video. For visualisizing audio we
compute spectral features from the audio samples which are subsequently mapped to
a 5-dimensional space. These five parameters are finally used for drawing parametric
shapes which are used as representative thumbnails. For videos we employ a story-
board based visualisation using indicative key frames.
Fig. 6. Prototype of the I-SEARCH result visualisation interface. http://vision.

iti.gr/sotiris/isearch/index.html
In any case we favor a hierarchical visualization of documents using three levels of

detail. At the first level single thumbnails are presented aiming at fast but crud
relevant judgments. The second level presents a more detailed view of the item both
in content and resolution. Finally the third stage involves downloading and
previewing the item in its original form. For documents containing several modalities
a stacking metaphor is used at the lowest level of detail, with the most relevant
modality on the top while for higher levels of detail the user may switch among
different modalities by means of a menu. If real world information is available, then
additional “views” are possible. Currently our system supports geographic
information (latitude-longitude coordinates) and temporal information (single
time-stamp for each document). This allows rearranging the document thumbnails to
reflect spatial or temporal relationships instead of document similarity.
7 Conclusions
A novel approach for multimodal search was presented in this article. I-SEARCH
allows easy retrieval of multiple media types simultaneously, namely 3D objects,
images, audio and video, using as queries combinations of different media types, text,
real-world information, expressions or emotions captured from the user with simple
devices. Several innovative solutions, which were developed within I-SEARCH,
constitute the proposed search and retrieval framework: a) a method for multimodal
descriptor extraction and indexing able to index COs irrespective of their constituting
modalities; b) a dynamic graphical use interface (GUI), enhanced with multimodal
querying capabilities; c) methods for analysing non-verbal emotional behaviour
expressed by full-body gestures and translating this behaviour to multimodal queries;
d) adaptive presentation of the search results using visual analytics technology. The
multimodal search engine is dynamically adapted to end-user’s devices, which vary
from a simple mobile phone to a high-performance PC. The framework will be
extended, including more functionalities, such as personalisation, relevance feedback,
annotation propagation and personalised recommendation exploiting social tagging.
The technologies implemented within I-SEARCH can potentially influence the FI
architecture and related frameworks. The outcomes of I-SEARCH can contribute to
Future Internet Public Private Partnership (FI-PPP) [19], which aims to advance
Europe's competitiveness in FI-related technologies and to support the emergence of
FI-enhanced applications of public and social relevance, more specifically to FI-
WARE Core Platform [18]. FI-WARE is expected to deliver an integrated service
infrastructure, building upon elements (called Generic Enablers) which offer reusable
and commonly shared functions making it easier to develop FI applications in
multiple sectors. Since multimedia/multimodal search has not yet been adopted by FI-
WARE, it can be proposed as a Generic Enabler of the FI-WARE core platform.
Acknowledgements. This work was supported by the EC-funded project I-SEARCH

(http://www.isearch-project.eu/isearch/).
References
1. Daras, P., Axenopoulos, A., Darlagiannis, V., Tzovaras, D., Le Bourdon, X., Joyeux, L.,
Verroust-Blondet, A., Croce, V., Steiner, T., Massari, A., Camurri, A., Morin, S.,
Mezaour, A.D., Sutton, L., Spiller, S.: Introducing a Unified Framework for Content
Object Description. International Journal of Multimedia Intelligence and Security, Special
Issue on Challenges in Scalable Context Aware Multimedia Computing 2(3-4), 351–375
(2011), doi:10.1504/IJMIS.2011.044765
2. Daras, P., Alvarez, F.: A Future Perspective on the 3D Media Internet. Towards the Future
Internet - A European Research Perspective (January 2009) ISBN: 978-1-60750-007-0
3. Champeon, S.: Progressive Enhancement and the Future of Web Design,
http://www.hesketh.com/thought-leadership/our-publications/
progressive-enhancement-and-future-web-design
4. Etzold, J., Brousseau, A., Grimm, P., Steiner, T.: Context-Aware Querying for Multimodal
Search Engines. In: Schoeffmann, K., Merialdo, B., Hauptmann, A.G., Ngo, C.-W.,
Andreopoulos, Y., Breiteneder, C. (eds.) MMM 2012. LNCS, vol. 7131, pp. 728–739.
5. Steiner, T., Sutton, L., Spiller, S., et al.: I-SEARCH – A Multimodal Search Engine based
on Rich Unified Content Description (RUCoD). Submitted to the European Projects Track
at the 21st International World Wide Web Conference. Under review,
http://www.lsi.upc.edu/~tsteiner/papers/2012/isearch-multi
modal-search-www2012.pdf
6. Zagoris, K., Arampatzis, A., Chatzichristofis, S.A.: www.MMRetrieval.Net: a multimodal
search engine. In: Proceedings of the Third International Conference on SImilarity Search
and APplications (SISAP 2010), pp. 117–118. ACM, New York (2010)
7. Bederson, B.B.: PhotoMesa: A Zoomable Image Browser Using Quantum Treemaps and
Bubblemaps.. In: ACM Symposium on User Interface Software and Technology, UIST
2001, CHI Letters, vol. 3(2), pp. 71–80 (2001)
8. Graham, A., Garcia-Molina, H., Paepcke, A., Winograd, T.: Time as essence for photo
browsing through personal digital libraries. In: Proceedings of the 2nd ACM/IEEE-CS
Joint Conference on Digital Libraries (JCDL 2002), pp. 326–335. ACM, NY (2002)
9. Platt, J.C., Czerwinski, M., Field, B.A.: PhotoTOC: automatic clustering for browsing
personal photographs. In: Proceedings of the 2003 Joint Conference and the Information,
Communications and Signal Processing Fourth Pacific Rim Conference on Multimedia,
December 15-18, vol. 1, pp. 6–10 (2003)
10. Zhang, H., Weng, J.: Measuring Multi-modality Similarities Via Subspace Learning for
Cross-Media Retrieval. In: Advances in Multimedia Information Processing, PCM (2006)
11. Yang, Y., Xu, D., Nie, F., Luo, J., Zhuang, Y.: Ranking with Local Regression and Global
Alignment for Cross Media Retrieval. ACM MM, Beijing, China (2009)
12. Axenopoulos, A., Manolopoulou, S., Daras, P.: Multimodal Search and Retrieval using
Manifold Learning and Query Formulation. In: ACM International Conference on 3D Web
Technology, Paris,France, June 20-22 (2011)
13. Daras, P., Manolopoulou, S., Axenopoulos, A.: Search and Retrieval of Rich Media
Objects Supporting Multiple Multimodal Queries. Accepted on IEEE Transactions on
Multimedia
14. Axenopoulos, A., Manolopoulou, S., Daras, P.: Optimizing Multimedia Retrieval Using
Multimodal Fusion and Relevance Feedback Techniques. In: Schoeffmann, K., Merialdo,
B., Hauptmann, A.G., Ngo, C.-W., Andreopoulos, Y., Breiteneder, C. (eds.) MMM 2012.
LNCS, vol. 7131, pp. 716–727. Springer, Heidelberg (2012)
15. Gennaro, C., Amato, G., Bolettieri, P., Savino, P.: An Approach to Content-Based Image
Retrieval Based on the Lucene Search Engine Library. In: Lalmas, M., Jose, J., Rauber, A.,
Sebastiani, F., Frommholz, I. (eds.) ECDL 2010. LNCS, vol. 6273, pp. 55–66. Springer,
Heidelberg (2010)
16. Glowinski, D., Dael, N., Camurri, A., Volpe, G., Mortillaro, M., Scherer, K.: Towards a
Minimal Representation of Affective Gestures. IEEE Transactions on Affective
Computing 2(2), 106–118 (2011)
17. Varni, G., Volpe, G., Camurri, A.: A System for Real-Time Multimodal Analysis of
Nonverbal Affective Social Interaction in User-Centric Media. IEEE Transactions on
Multimedia 12(6), 576–590 (2010)
18. http://www.fi-ware.eu/
19. http://www.fi-ppp.eu/
Semantically Enriched Services
to Understand the Need of Entities
Flávio de Oliveira Silva1,3 , Alex Dias2 , Caio César Ferreira3 ,

Eduardo De Souza Santos3 , Fabı́ola Souza Fernandes Pereira4,
Isabelle Cecı́lia de Andrade3 , João Henrique de Souza Pereira1,
Lásaro Jonas Camargos3, Luiz Cláudio Theodoro3,
Maurı́cio Amaral Gonçalves3 , Rafael Pasquini3 , Augusto José Venâncio Neto5 ,
Pedro Frosi Rosa3 , and Sergio Takeo Kofuji1
1
University of São Paulo, Brazil
{flavio,kofuji}@pad.lsi.usp.br, [email protected]
2
Federal Institute of Triângulo Mineiro, Brazil
[email protected]
3
Federal University of Uberlândia, Brazil
{caiocf,isabelle,mauricio}@algartelecom.com.br, [email protected],
{lasaro,pasquini,frosi}@facom.ufu.br, [email protected]
4
Algar Telecom, Brazil
[email protected]
5
Federal University of Ceará, Brazil
[email protected]
Abstract. Researchers from all over the world are engaged in the de-
sign of a new Internet, and Software-Defined Networking (SDN) is one
of the results of this engagement. Net-Ontology uses a SDN approach
to bring semantics to the intermediate network layers and make them
capable of handling application requirements and adapt their behaviour
over time as required. In this paper we present an experimental evalu-
ation of Net-Ontology and a feature comparison against the traditional
TCP/IP stack. This paper extends our earlier work towards a Future
Internet, showing a viable approach to introduce semantics at network
lower layers by contributing to bring richer and efficient services.
Keywords: Future Internet, Enrich Services, Network Ontology, SDN,

DTS, Workspace.
Introduction
The evolution of the intermediate network layers have been lagging behind that
of the lower and upper layers. The Internet Protocols, specified more than three
decades ago, are the likely culprit; the application needs have changed by leaps,
while the TCP/IP has only been patched, trying to meet these requirements.
Over the last few years, the networking community has strived to correct this
phenomenon[1, 3, 4, 21].

Semantically Enriched Services to Understand the Need of Entities 143
Researchers from all over the world are engaged in the design of a new Internet,
from the ground up. This so called clean slate approach, frees the research from
the legacy of the current architecture and fosters innovations[18]. At a future
time, when results should be deployed, the research will then be refocused to the
transition from the current Internet to the future Internet
One of the results of this effort to create the Future Internet is Software-Defined
Networking (SDN)[5, 6]. SDN enables researchers to innovate and experiment new
network protocols, naming and addressing schemes, such as the one presented in
this paper, which aims at bridging the evolutionary gap between upper, lower, and
the intermediate network layers by using a richer semantics [15, 16].
FINLAN (Fast Integration of Network Layers) [9, 13, 14, 19] aims at provid-
ing high adaptability through the use of semantic concepts based on ontology,
with the elimination of static routing and addressing tied to physical location,
resulting in a better and efficient utilization of the network infrastructure.
FINLAN defines two intermediate layers that communicate between each
other using OWL (Web Ontology Language), but that clearly differentiate in
function: DL-Ontology and Net-Ontology.
The DL-Ontology layer is essentially responsible for data transfer between
the Physical layer and the upper layers, handling the semantic communication
between the peer entities and bringing a richer capacity to express their require-
ments. On the other hand, the Net-Ontology layer is accountable for handling
service needs, as it is capable of understanding specific requirements from the ap-
plication and adapting the communication to support them only when required,
using DL-Ontology to deal with the semantic communication.
In this chapter we present the Net-Ontology layer, which sits between the
DL-Ontology layer and the application. We also present its implementation and
a first experimental evaluation. The implementation presented is based on the
Title Model[17], our vision regarding future networks.
The remainder of this work is organized as follows: Section 1 describes the
Net-Ontology. Section 2 shows the Net-Ontology implementation and Section 3
the experimental results. The conclusions are presented in Section 4.
1 The Net-Ontology
The DL-Ontology is the lower layer of the FINLAN stack depicted in Figure
1, and enables the communication using concepts expressed in OWL over the
Physical layer.
The Net-Ontology layer is responsible for supporting the service needs of the
upper layer and deliver them to the DL-Ontology layer, built according to the
FINLAN Ontology. In this approach, the Net-Ontology is able to understand
specific requirements of a given application that may arise over communication
and provide them.
For example, let us suppose that two persons, P1 and P2 , are chatting over
the Internet, using the application FinChat that runs over the FINLAN stack.
In a certain moment, they want to start a secret conversation. To FinChat meet
144 F. de Oliveira Silva et al.
this need, the only thing it has to do is to inform the Net-Ontology layer that
from now on the chat is to be confidential. The Net-Ontology layer is able to
understand this need and act accordingly modifying all packets exchanged from
that moment.
The Net-Ontology consists, basically, of two main modules: requirement anal-
ysis and requirement manager, as depicted in Figure 1.
Fig. 1. Net-Ontology modules
The requirement analysis module (RAM) is responsible for handling the ap-
plication requests regarding communication requirements. To accomplish this,
RAM uses the Leśniewski’s logic as proposed in [8]. The purpose is to manage
the services requirements over time. This module recognizes what technologi-
cal features are necessary to satisfy a given requirement, in a given moment,
combining them in logical formulas.
As an example, let us suppose that a service S1, in a moment t1 , may need
to establish communication with the service S2, with a specific requirement.
The RAM will verify that this upper layer requirement can be provided by the
technological requirements R1 and R2. In another moment t2 , S1 wishes to
improve security, using confidentiality in the conversation. For so, it is necessary
the technological requirement R3. These scenarios will be interpreted by the
analysis module and represented by the following axioms:
S1S2t1 → R1 ∧ R2 (1)
S1S2t2 → (R1 ∧ R2) ∧ R3 (2)
The requirement manager module (RMM) takes the rules requirements and
transform them into FINLAN ontology fragments. Besides that, this module
is able to interpret and deploy the algorithms correlated with each requirement
of the ontology in the network stack.
Taking the aforementioned example, in the moment t1 , RMM receives the

requirements R1 and R2 from RAM. It will then use the FINLAN Ontology, and
add to the packages an OWL fragment, representing that P1 has requirements
R1 and R2 in FinChat.
From now on, the packages will be transmitted containing new information.
When the FinChat of P2 receives an OWL package, meaning that R1 and R2
are required, the RMM will be able to understand and make use of the necessary
algorithms. The intelligence for the network to understand and implement the
applications needs is the main responsibility of the requirement manager module.
The requirements, manipulated by RAM, are stored at the Domain Title Ser-
vice (DTS), which consists of a distributed system over the network elements
responsible for maintaining the entities available in a domain and their commu-
nication requirements over time. It plays an important role at central aspects of
networking like naming and addressing, and has the ability to share the context
among communicating entities. This sharing is provided by the workspace.
The workspace is a logical bus which contains network elements required to
support the communication of the entities. The workspace is created by an entity
wiling to communicate with a specific purpose and thus defines its requirements
and capabilities. A new entity can be joined to an existing workspace and, in
such event, the logical bus can be adapted to handle its communication.
All entities that shares a workspace takes part in the same communication.
The data is sent once by a source to the workspace and is received by all the
others, thus making an efficient use of the physical layer.
In the next subsection it will be presented a complete case of how the Net-
Ontology modules interact with the others FINLAN layers and the DTS.
1.1 FINLAN Semantic Communication
The communication between the FINLAN layers occurs in a semantic way, by

using OWL. Below, it is presented an example to illustrate how this communi-
cation happens.
Let us suppose a scenario where John and Paul are chatting using the applica-
tion FinChat that runs over the FINLAN layers, through the workspace WKS.1.
In a first moment t1 , they are just talking about irrelevant issues and are not
concerned about any additional feature that FinChat can offer to them. So, the
packages travelling in the network are very simple, and the Net-Ontology has
not introduced any new requirement at the communication, in this case, only
the DL-Ontology handles their communication. A code snippet example can be:
< Message rdf:ID = " M e s s a g e _ 1" >
< w o r k s p a c e I D r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " >
WKS .1 </ w o r k s p a c e I D>
< source r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > Paul </
source >
< d e s t i n a t i o n r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " >
John </ d e s t i n a t i o n>
< payload r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# byte " > Hello !
How are you doing ?? </ payload >
</ Message >
After some time, at instant t2 , John starts an important subject, and selects
the feature delivery guarantee of FinChat. This means that from now on, FinChat
requires delivery guarantee to the network. The Figure 2 shows the messages
flow that will be sent and received between the FinChat entities and the DTS,
to attend this request.
With a new requirement, the Net-Ontology layer is triggered, and the require-
ment analysis module checks that it is necessary the technological requirement
of a delivery guarantee algorithm. John’s FinChat, then, sends the following
control message to DTS:
< C o n t r o l M e s s a g e rdf:ID = " C o n t r o l M e s s a g e _ 1" >
< A p p l i c a t i o n rdf:ID = " FinChat " >
< HasNeed >
< D e l i v e r y G u a r a n t e e rdf:ID = " D e l i v e r y G u a r a n t e e _ 0 1" / >
</ HasNeed >
</ A p p l i c a t i o n>
< source r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > John </
source >
DTS </ d e s t i n a t i o n>
< payload r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " >
AddNeed </ payload >
</ C o n t r o l M e s s a g e>
After registering John’s need, the DTS will send him a confirmation message:
< C o n t r o l M e s s a g e rdf:ID = " C o n t r o l M e s s a g e _ 1 R" >
< A p p l i c a t i o n rdf:ID = " FinChat " / >
< source r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > DTS </
source >
John </ d e s t i n a t i o n>
< payload r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > OK </
payload >
Fig. 2. Message flow example for a new requirement
At the same time, DTS will also send to Paul, who is in the same workspace as
John, a control message, asking if the need requested is supported:

< HasNeed >
</ HasNeed >
source >
Paul </ d e s t i n a t i o n>
< payload r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > isOK <
/ payload >
If Paul ’s FinChat can supply the delivery guarantee feature, the response below
is sent to DTS and it is established a communication with support to delivery
guarantee:
< A p p l i c a t i o n rdf:ID = " FinChat " / >
< source r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > Paul </
source >
payload >
In case of Paul ’s FinChat with no support for the delivery guarantee, this feature
will not be present in the communication between both applications.
Notice that through the Net-Ontology, FINLAN is able to register the services
needs into the DTS. From now on, it can manage what is the best way to deliver
FINLAN packages.
If a third person, Ringo, wants to join the conversation, Ringo’s FinChat will
handshake with DTS to check if it has support to DeliveryGuarantee 01. This
scenario is illustrated in Figure 3.
Fig. 3. Message flow example for a Join into Workspace WKS.1
The following messages are exchanged and Ringo joins the workspace WKS.1.
After the joining and, hence, sharing of the workspace, Ringo’s FinChat and
all the other entities will receive the same data messages without the need of
multiple data flows.

< source r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > Ringo <
/ source >
< payload r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > Join <
/ payload >

< HasNeed >
</ HasNeed >
source >
Ringo </ d e s t i n a t i o n>
< payload r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > isOK <
/ payload >

< A p p l i c a t i o n rdf:ID = " T e s t A p p l i c a t i o n" >
< source r d f : d a t a t y p e= " http: // www . w3 . org /2001/ X M L S c h e m a# string " > Ringo <
/ source >
payload >
It is also important to mention that after the exemplified handshakes between

DTS and FinChat entities, the requirement manager module is responsible for
guaranteeing that the exchanged packages during the chat will have the nec-
essary information, implementing the algorithm DeliveryGuarantee 01. For ex-
ample, differently from the Message 1 structure, the messages must have an
identification field, through which the control of lost packages is made.
2 Implementation
Our FINLAN stack implementation consists of a Java library that uses commu-
nication interfaces through Raw Sockets. The linking between Java and C por-
tions of the code was done in Java Native Interface (JNI) [16, 19], as depicted in
Figure 4.
It is observed that the application App.java should use the API available in
the library Finlan.jar to establish communication. In this way, when an appli-
cation sends a packet, it communicates with the Net-Ontology sending its char-
acteristics. According to these characteristics, the Requirement Module Analisys
determines, through an inference engine, the application needs and proceeds
Fig. 4. FINLAN Modules Implementation
with the delivery of these. After the completion of the relevant operations, Net-
Ontology sends the primitive to the DL-Ontology which, in turn, takes care of
sending the packet through the JNI interface to the libFinlan.so library.
3 Experimental Results
To validate the implementation of this work, it was performed experiments com-

paring FINLAN with the TCP/IP Architecture protocols. The goal of these
experiments was to show the behavior of the ontology use in a file transfer op-
eration with the delivery guarantee need activated, illustrating the use of the
Net-Ontology layer in FINLAN.
The TCP/IP protocols, by definition, already implements the delivery guaran-
tee feature, when the TCP transport protocol is used. To FINLAN support this
need, this work implemented it in the Net-Ontology layer, using the algorithm
described in [14]. This algorithm is a mechanism to ensure that all packages sent
are received and works as follows: when the need delivery guarantee is activated
by the application layer, all packages, sent from this moment contains a new
field representing the number of the package.
In parallel, there is a mechanism of confirmation requests and responses mes-
sages: the source host informs the packages already sent, requesting the lost ones.
The destination, on the other hand, answers which packages it did not receive.
This change of confirmation messages is orchestrated by the RTT variable, pro-
posed in [7], which consists of the best estimate (for that moment) for the send
and receive time up of the packets destination.
3.1 Network Traffic Evaluation
The experiments were performed over the following environment: hosts with 4GB
of RAM, CPU Intel R
CoreTM 2 DUO @ 2.10GHz, running Linux operational
system with kernel 2.6.41.10-3.fc15.x86 64. The files transfered have size of 1, 5,
10, 15, 20, 25, 30, 35, 40, 45 and 50MB. The RTT variable was set to a fixed
(a) Number of packets
(b) Percentual reduction
Fig. 5. FINLAN and TCP traffic comparison
value of 1 second. Figure 5 shows the results, comparing the number of packets
transmitted in both: FINLAN and TCP.
It is possible to observe that in the scenarios of this experimentation, FINLAN
had a smaller number of transmitted packets. In the transfer operation of 10MB,
for example, FINLAN transmitted 8140 packets, while the TCP transmitted
10631 (one difference of 30.6 percent).
This is due to the delivery guarantee algorithm implemented in FINLAN that
sends confirmation messages in intervals of the RTT, informing only the lost
ones, in a period, to be re-transmitted, while TCP transmitted several number of
ACK packages. This confirms that the network traffic packets is decreased using
the delivery guarantee algorithm implemented over a stack that semantically
understands the concepts and adapts the messages from this understanding.
To illustrate the primitives in these experiments, Figure 6 shows snapshots
from the Wireshark of two packets captured during the transmission of the 50MB
file. The first one, in Figure 6(a), is the confirmation request of the source entity,
called “fabiola”, informing that the range of packages from 133 to 367 was sent.
The Figure 6(b) represents the response, confirming the lost packages, through
(a) Confirmation request
(b) Confirmation response
Fig. 6. Snapshots of FINLAN confirmation messages
the field LostMessageQuantity. According to this capture, the packages from 220
to 281 were lost and only them were re-transmitted.
4 Conclusions
This work presented the Net-Ontology Layer, experimental results of its imple-
mentation and how it is possible to use ontology at the intermediate networks
layers to understand and support different entities needs.
The results of using ontology to support the delivery guarantee need demon-
strate an optimization of more than 30 percent of the packets sent in a file
transfer, compared with the traditional TCP/IP protocols usage.
By the Net-Ontology use, it was demonstrated the possibility to substitute
the traditional TCP/IP protocols used at the transport and network layers. This
brings more semantic power for the Future Internet networks, as the network
intermediate layers become able to better understand the entities needs.
Future Internet is being constructed with worldwide collaboration and is based
on research and experimentation. Our previous work showed [17, 19] how FIN-
LAN approach and the Title Model Ontology can work together with different
efforts regarding the future, while the work presented details on how these pro-
posals can come true.
As future works, it is expected to experiment the Net-Ontology implemen-
tation in different testbeds, such as OFELIA [11] and FIBRE (Future Internet
testbeds/experimentation between BRazil and Europe)[2, 20]. In complement,
it will be finished the actual working in progress to the experimentation using
OpenFlow [10]. Also, experimental tests using workspaces for multicast aggre-
gation [12] are being executed at OFELIA testbed.
The research and experimentation results show that we are facing a viable
approach to introduce semantics at network lower layers, by contributing to
bring richer and efficient services.
Acknowledgment. This work received contributions from the FINLAN project

and other researchers from the MEHAR program. The authors would like to
thank all FINLAN and MEHAR members for the extensive discussion and col-
laboration to the Future Internet perspectives.
credited.
References
[1] Clayman, S., Galis, A., Chapman, C., Toffetti, G., Rodero-Merino, L., Vaquero,
L.M., et al.: Monitoring Service Clouds in the Future Internet. Towards the Future
Internet - Emerging Trends from European Research, p. 115 (2010)
[2] FIBRE: FIBRE Project (January 2012), http://www.fibre-ict.eu/
[3] FIRE: FIRE White Paper (August 2009), http://www.ict-fireworks.eu/
fileadmin/documents/FIRE White Paper 2009 v3.1.pdf
[4] Galis, A., Denazis, S., Bassi, A., Giacomin, P., Berl, A., Fischer, A., et al.: Manage-
ment Architecture and Systems for Future Internet. Towards the Future Internet
- A European Research Perspective, p. 112 (2009)
[5] Greenberg, A., Hjalmtysson, G., Maltz, D.A., Myers, A., Rexford, J., Xie, G.,
Yan, H., Zhan, J., Zhang, H.: A clean slate 4d approach to network control and
management. SIGCOMM Comput. Commun. Rev. 35, 41–54 (2005),
http://doi.acm.org/10.1145/1096536.1096541
[6] Greene, K.: TR10: Software-Defined Networking. MIT - Technology Review (2009)
[7] Jacobson, V.: Congestion Avoidance and Control. SIGCOMM Communications
Architectures and Protocols, USA 88, 314–329 (1988)
[8] Leśniewski, S.: Comptes rendus des séances de la Société des Sciences et des Lettres
de Varsovie. Class III, pp. 111–132 (1930)
[9] Malva, G.R., Dias, E.C., Oliveira, B.C., Pereira, J.H.S., Kofuji, S.T., Rosa, P.F.:
Implementação do Protocolo FINLAN. In: 8th International Information and
Telecommunication Technologies Symposium (2009)
[10] McKeown, N., Anderson, T., Balakrishnan, H., Parulkar, G., Peterson, L., Rex-
ford, J., Shenker, S., Turner, J.: OpenFlow: Enabling Innovation in Campus Net-
works. SIGCOMM Comput. Commun. Rev. 38, 69–74 (2008)
[11] OFELIA: OFELIA Project (January 2012), http://www.fp7-ofelia.eu/
[12] de Oliveira Silva, F.: Experimenting domain title service to meet mobility and
multicast aggregation by using openflow. In: MyFIRE Workshop (September 2011)
[13] Pereira, F.S.F., Santos, E.S., Pereira, J.H.S., Rosa, P.F., Kofuji, S.T.: Proposal
for Hybrid Communication in Local Networks. In: 8th International Information
and Telecommunication Technologies Symposium (2009)
[14] Pereira, F.S.F., Santos, E.S., Pereira, J.H.S., Rosa, P.F., Kofuji, S.T.: FINLAN
Packet Delivery Proposal in a Next Generation Internet. In: IEEE International
Conference on Networking and Services, p. 32 (2010)
[15] Pereira, J.H.S., Pereira, F.S.F., Santos, E.S., Rosa, P.F., Kofuji, S.T.: Horizontal
Address by Title in the Internet Architecture. In: 8th International Information
and Telecommunication Technologies Symposium (2009)
[16] Pereira, J.H.S., Santos, E.S., Pereira, F.S.F., Rosa, P.F., Kofuji, S.T.: Layers Op-
timization Proposal in a Post-IP Network. International Journal on Advances in
Networks and Services (2011)
[17] Pereira, J.H.S., Silva, F.O., Filho, E.L., Kofuji, S.T., Rosa, P.: The Title Model
Ontology for Future Internet Networks. In: Domingue, J., et al. (eds.) FIA 2011.
[18] Roberts, J.: The clean-slate approach to future internet design: a survey of research
initiatives. annals of telecommunications - annales des télécommunications 64(5-6),
271–276 (2009), http://www.springerlink.com/content/e240776641607136/
[19] Santos, E., Pereira, F., Pereira, J.H.: Meeting Services and Networks in the Future
Internet. In: Domingue, J., et al. (eds.) FIA 2011. LNCS, vol. 6656, pp. 339–350.
[20] Stanton, M.: FIBRE-EU and FIBRE-BR (October 2011), http://www.future-
internet.eu/fileadmin/documents/poznan documents/Session2 3 Inter
national coop/2-3-stanton.pdf
[21] Tselentis, G., Domingue, J., Galis, A., Gavras, G., Hausheer, D., Krco, S., Lotz,
V., Zahariadis, T.: Towards the future internet a European research perspective.
IOS Press, Amsterdam (2009)
Supporting Content, Context and User Awareness
in Future Internet Applications
Paweł Świątek, Krzysztof Juszczyszyn, Krzysztof Brzostowski,

Jarosław Drapała, and Adam Grzech
Institute of Computer Science, Wrocław University of Technology

Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, Poland
{pawel.swiatek,krzysztof.juszczyszyn,krzysztof.brzostowski,
jaroslaw.drapala,adam.grzech}@pwr.wroc.pl
Abstract. We show how the applications utilizing a Future Internet architecture

can benefit from its features like quality of service (QoS) provisioning and
resources reservation. We demonstrate, how proposed applications address
content, context and user awareness basing on the underlying Next Generation
Network (NGN) infrastructure and how it can be used to host service-based
applications.
Keywords: Future Internet, Next Generation Networks (NGN), distributed

applications, quality of service (QoS), content-awareness, context-awareness,
services personalization, service oriented architecture (SOA).
1 Introduction
One of the main motivations for designing new architectures for the Future Internet is
to meet challenges imposed on the ICT infrastructure by new applications. These
challenges include among others:
1. Content awareness – meaning the sensitivity of data processing and transmission
methods to the content being delivered to the end-user. Content awareness may
emerge in: different processing of various data streams (i.e. video encoding or
sensor data encryption) and different forwarding methods (e.g. routing) for various
streams.
2. Context awareness consisting in different treatment (in terms of forwarding and
processing methods) of traffic depending on the particular use-case scenario of
application generating this traffic. Context may be connected for example with the
type of networking device used by a user or users geographical localization.
3. User awareness understood as personalization of services delivered to end-user.
Personalization is achieved by means of proper choice of data processing and
transmission methods according to functional and non-functional requirements
stated by the user. Users requirements may be formulated explicitly or be a result
of automatic recommendation which is based on the history of the application
usage.
Supporting Content, Context and User Awareness in Future Internet Applications 155
4. Sensor networks and applications covering such applications such as: smart energy
metering, vehicle networks, intelligent building infrastructure, telemedicine, etc.
Each particular telemetry application involves specific types of data processing
methods and transmission of large number of small portions of data often requiring
real-time or near real-time end-to-end performance.
Augmentation of the current Internet architecture with the abovementioned

functionalities will fulfill the assumptions of the pervasive computing paradigm where
end-to-end services delivery is facilitated by a cloud of distributed networking devices
and loosely coupled application modules. The key feature of such an approach is the
user-centricity where the user does not invoke any particular applications or service
nor even specifies where the application should be executed.
Currently there is a number of approaches aiming at meeting the assumptions of
the Future Internet. Depending on the proposed ICT architecture some or all of them
are assured by utilization of proper networking techniques and concepts. These
approaches differ among other in the layer in which the new Future Internet
functionalities are accomplished. As an example consider Content Centric Network
(CCN) proposed by Van Jacobson [10] where content delivery mechanism (e.g.:
caching, forwarding, security, etc.) are mostly implemented at lower network layers.
This revolutionary post-IP approach requires the entire networking protocol stack to
be redesigned. On the other hand in the prototype of the Parallel Internet CAN
(Content-Aware Network) [2] being one of the results of the Polish national project
“IIP” (polish acronym for Future Internet Engineering) [4] the content is encapsulated
in new frames format, but signaling messages are passed with use of the IPv6
(Internet Protocol version 6) protocol and content delivery routes calculation and
caching is accomplished at the application layer. These two approaches are candidates
for implementing the concepts of the internet of content and media.
One of the most mature architectures for the Future Internet is the Next Generation
Network (NGN) [9]. The NGN signaling system in conjunction with the
Differentiated Services (DiffServ) quality of service assurance model [3] and IPv6
networking protocol stack allows for implementation of converged all-IP multi-
domain network conforming with all assumptions of the Future Internet. Sample
implementation of this approach is the Parallel Internet IPv6 QoS prototype [15,16].
The main contribution of this work is the concept of how to achieve content,
context and user awareness in the IPv6 QoS architecture by proper signalization in the
service and transport stratum of the NGN architecture. The proposed concept is
illustrated on exemplary applications designed for the IPv6 QoS system. In section 2
we give a brief overview of the architecture of the IPv6 QoS system with special
focus put on the service stratum signaling system in section 2.2. Next, in section 3 we
present exemplary applications and show how content, context and user awareness is
achieved with use of service stratum signaling. Additionally, we show in section 3.3
how custom personalized application can be designed with use of processing and
communication services already existing in the system. In section 4 we conclude our
work and point out directions of future works.
156 P. Świątek et al.
2 Systems Architecture
2.1 IPv6 QoS System

In this work we consider an IPv6 QoS system architecture developed in the polish
national project IIP [4]. In this architecture it is assumed that the system consists of
multiple layers each of which provides certain functionalities to the adjacent upper
layer. The first layer is a physical network infrastructure which with use of
virtualization techniques [7] provides to the second layer virtualized networking
environment with dedicated communication and processing resources. Such
virtualization allows for coexistence of multiple isolated virtual networks (called
parallel internets - PI), characterized among others by different frame formats,
protocol stacks and forwarding capabilities, in a single physical infrastructure.
IPv6 QoS system is one of parallel internets existing in a virtual networking
environment. In general the IPv6 QoS architecture is based on coupling of the DiffServ
quality of service assurance model and NGN signaling system. DiffServ is responsible
for delivery to traffic flows generated by users required level of the quality of services
by means of flow admission control, classification of flows to predefined traffic classes
and processing of aggregated flows from different traffic classes. The NGN signaling
system is used to provide end-to-end QoS guaranties by reserving necessary amount of
communication resources to each particular connection request. Reservation of
communication resources is performed by assignment of the request to proper DiffServ
traffic class, which meets the QoS requirements for this flow.
The purpose of signaling in NGN is twofold. The first one is to reserve required
communication resources and to establish an end-to-end connection between a pair of
hosts in the system. This signaling is performed at the network layer in so-called
transport stratum. Second type of signaling is performed at the application layer
(service stratum). Service stratum signaling is in general an application specific
signaling (e.g. SIP signaling) the aim of which is to configure distributed modules of
an application and to process information necessary to send to transport stratum a
request for communication resources reservation. Signaling can be also viewed as a
middleware which separates the networking layer functionalities and application
domain-specific specific functionalities.
2.2 Service Stratum Signaling

The task of service stratum signaling is to control the execution of distributed
applications and to pass communication resources reservation requests from
applications to the network. Service stratum being an intermediate layer between
applications and the network which translates application specific signaling and
negotiations to uniform service stratum – transport stratum interface allows for
implementation of arbitrary application specific signaling schemes. This in turn
allows achieving content, context and user awareness by implementation of
specialized services’ management mechanisms, whose task is to transparently
compose and control execution of personalized complex services based on functional
and non-functional requirements.
In our approach, based on the Service Oriented Architecture (SOA) paradigm, we

assume that applications in the IPv6 QoS system consists of distributed loosely
coupled modules (called atomic services). Execution of each application use-case
scenario is performed by sending a request to an application server, which composes a
specialized complex service from the available in the system atomic services.
Additionally, we assume that end-to-end communication services provided by the
IPv6 QoS system are also treated as atomic services and can be utilized to deliver to
the user requested complex services according to SOA approach.
In order to deliver to the users requested complex services a two-stage signaling in
service stratum is proposed. The task of the first stage of signaling is twofold. Firstly,
based on services available in distributed communication system, it allows to
compose a complex service which conforms with functional and nonfunctional
requirements [8]. Secondly, it notifies each module of distributed application how and
with which module they should establish communication in order to deliver requested
complex service. The aim of the second stage signaling is to negotiate the details of
communication between each pair of atomic services taking part in complex service
delivery. Communication details depend on the functionalities of communicating
services and may concern among others: data formats, audio and video codec,
required transport protocol, encryption, etc. Taking into account negotiated
communication details and non-functional requirements concerning requested
complex service proper end-to-end communication service is requested from the IPv6
QoS system for each pair of communicating atomic services. Note that thanks to the
negotiation process requested end-to-end communication services depend on the
context of communication and the content being transmitted resulting in fully
personalized context and content aware complex services delivery.
Fig. 1. Signaling messages exchange for an exemplary complex service
Fig. 1. Signaling messages exchange for an exemplary complex service. an

exchange of signaling messages required to prepare an exemplary service and
consisting in sending requested data from a service to a user is presented. In the first
stage an arriving service request (1) is processed by server (2) in order to compose a
complex service conforming with users requirements. The result of a composition

process is a set of atomic services available in distributed system which will take part
in execution of request (1). In the second stage server configures all necessary atomic
services (including client application) by sending them information about: source of
data, data processing methods and destination of processed data (3a and 3b). In other
words in this stage each service is notified from who and what will receive, what to
do with received data and where to send it. After configuration is confirmed to the
server (4a and 4b) each pair of communicating services negotiates values of
application and communication specific parameters (5 and 6). When communicating
parties agree upon type of communication appropriate end-to-end communication
services (which guarantee delivery of required QoS parameters for data transfer) it is
requested from the IPv6 QoS system. This is accomplished by sending a resources
reservation request to the resources and admission control function (RACF) (7) with
use of service control function (SCF). After confirmation of connection preparation
(8) data transmission may take place (9).
The first stage of signaling beginning with sending of request (1) and ending with
configuration of all services (4a and 4b) is accomplished with use of XML-RPC
protocol. The second stage of signaling consisting of negotiation of values of
communication and other application specific parameters (5 and 6) is accomplished
with use of XMPP protocol. It is important to note, that vertical communication
(signaling) of application components and network is done with use of service
stratum – transport stratum interface (SCF – RACF to be exact). This means that each
application component that is able to send requests to the network should be equipped
with SCF module which translates application specific horizontal signaling to
application independent uniform vertical signaling between service and transport
stratum defined by SCF-RACF interface.
3 Future Internet Applications
3.1 SmartFit
Sustained progress and developing of infrastructure for wireless sensor networks and
wearable sensors makes basis for pervasive computing systems. Such systems can
operate in distributed environment where wireless sensor networks consist of huge
amount of low-cost, low-power sensing nodes and many different services for data
transfer, processing, storage and supporting decision making [1]. Sensing nodes can
be used to build sensor networks such as Body Area Networks (BAN) or Personal
Area Networks (PAN). On the other hand we have vast number of services in
distributed environment facilitating the access to one or more functionalities.
SmartFit is a system adopting new technologies of pervasive computing and was
designed to support endurance and technical training of either amateur and elite
athletes. Application such as SmartFit must be designed to provide its functionalities
“anywhere and anytime”. It means that acquired data must be transmitted between
users of the system (i.e. athletes and trainers) with predefined quality level
independently of their location. In order to fulfil this requirement each functionality
was decomposed on small modules called atomic services. For each atomic service
few different required levels of quality was defined. It means that we have different
versions of each atomic service. These different versions of atomic services are used
in the process of user-centric functionality delivery with use of orchestration
mechanism. User-centric functionality means that in order to compose such
functionality user’s specific requirements and needs are taken into account.
In Fig. 2 general architecture of SmartFit is presented. The first tier is used to
sensor data acquisition. The second tier is data processing and decision making tier.
The last one is presentation tier. For each tier the set of atomic services is defined. In
the process of user-centric functionality composition all versions of atomic services at
each tier are taken into account.
Fig. 2. Architecture of SmartFit system for distributed environment
One of the main functionality of the system is planning volume of endurance

training. Because in all sports endurance training is required, this functionality is
universal and can be used in the same way by athletes in various sports. The other
functionality of the system is connected with endurance training monitoring. It is used
to supervise correctness of the performed exercises, their right order and number of
repetitions. Moreover, such functionality can be used to predict injury.
Another functionality of SmartFit system is designed to support technical training.
On the contrary to endurance training, in this case, specific requirement for particular
sport must be taken into account. Fortunately, architecture of SmartFit system
provides mechanism for adding new functionalities supporting technical training of
different sports easily. For an example use-case scenario of the SmartFit system a
service for supporting a technical training in tennis was designed and implemented.
In Fig. 3 network topology configured to support technical training is presented. The
main element of this network is SmartFit server which supervise the whole process of
functionality delivery. This process has three phases. In the first of them the server must
configure each necessary atomic service which was planned to be used in particular
scenario. In considered example such atomic services facilitate physical and kinematic
data acquisition from user’s BAN, change detection and feature extraction for acquired
data, classification based on results of feature extraction, modelling of human
movements. The second phase is related to signalization between distributed modules
with use of SIP and XMPP protocols. The last phase of the functionality delivery
process is data transmission, processing with use of appropriate atomic services and
presentation of obtained results.
The above-mentioned atomic services can be used to compose functionality

supporting skill assessment and skill improvement of the elementary tennis strokes
such as serve, backhand and forehand. Skill assessment allows to determine current
skill level of tennis player. It can be helpful to make recommendations related to
support planning of future technical training of elementary tennis strokes.
Additionally, SmartFit is designed to support feedback training. It means that it is
possible to see, in real-time, physiological and kinematic data from sensors placed on
athlete’s body during stroke performance. This feature allows tennis player and/or
trainer to compare his performance to reference data acquired from his past trainings
or high level elite tennis players.
Fig. 3. Network topology supporting technical training scenario
Let us investigate the process of technical training of tennis player in details. In

this use-case scenario for SmartFit athlete’s BAN consists EMG, gyroscopes and
accelerometers wireless units. Acquired physiological and kinematic data is
transferred to the server. In skill assessment case signals from gyroscopes and
accelerometers are processed in order to build relationship between wrist flexion,
upper arm rotation and racquet speed. Obtained results can be compared with results
captured from reference data of high level elite tennis player. Based on this data it is
possible to build personalised model of improvement for a tennis player. It may be
also used to make recommendation for future technical training.
In the second case feedback and learning by example trainings are combined in
order to improve results of technical training. To this end acquired data from EMG,
gyroscopes and accelerometers units are transferred to SmartFit server. EMG data is
filtered in order to present to the trainer the sequence of muscle activation of upper
limb during strokes. Whereas signals from gyroscopes and accelerometers are used to
estimate trajectory of upper limb movement during tennis strokes such as serve,
forehand and backhand. In this case parameters of upper limb movement’s model
must be determined with use of delivered signals from gyroscope and accelerometers
units. Finally, the results of trajectory estimation is visualised and delivery to the
trainer and/or athlete.
In order to provide the required by the user quality of service it is necessary to
apply mechanism allowing for context awareness. Context awareness incorporated in
SmartFit can be used to adapt the packet size according to the user’s requirements.
Context information can be obtained through sensor networks e.g. measurement of
heart rate during training session or/and personal server e.g. GPS or networking
devices. Based on this information it is possible to predict user behaviour and his/her
location. Such mechanism facilitating SmartFit system with functionalities for
efficient management of network and computational resources in order to deliver
system functionality with required by user’s quality of service.
3.2 OnLine Lab

Virtual laboratory infrastructure should automate most tasks related to the reliable and
reproducible execution of required computations [12,17]. The application Online Lab
is a distributed, service-based computational laboratory benefiting from the IPv6 QoS
architecture which is used to distribute computational tasks while maintaining the
quality of service and user experience [13]. It allows its users, i.e. students or
researchers, to access different kinds of mathematical software via Python math
libraries and perform computations on the provided hardware, without the need for
installing and configuring any software on a local computer. The communication
mechanisms are designed for optimization of the users’ Quality of Experience,
measured by the response delay. The functionality of Online Lab embraces:
• access to computational services ensured by user’s virtual desktop which is

windowed interface opened in a Web browser,
• creation and removal of computational services with no limitations being
assumed on the nature of computations – the users may freely program
computational tasks in any language interpreted by running computational
services,
• user profile maintenance and analysis – the users are distinguished by their
profiles which hold information about their typical tasks and resource
consumption.
Online Lab (OL) implements an architecture consisting of user interface (OL-UI),
core server (OL-CORE), services and computational engines (OL-Services, based on
the Python engine in the current prototype). OL-UI is a web service emulating a
desktop and a window manager. Code is being typed into specialized data spaces -
notebooks, which are executable documents executed by OL-Services.
The process of user’s query execution is presented in Fig. 4. OL-Core and OL-
Services belong to the service. One notebook represents one computational task. The
system also may recommend notebooks of other users. The content of the notebooks
is annotated with the help of domain (Math) ontology.
Fig. 4. The general schema of the Online Lab service execution
OL-Core is constantly monitoring the OL-Services, storing execution times and data
transfers in its database. From the user point of view, in the case of computational tools,
the key element of the Quality of Experience (QoE) is waiting time. The waiting time is
the sum of computation time and communication times. The first is query-specific and
must be taken into account as a value predicted on the basis of known history of user
queries. The second depends on the volume of data and code.
Online Lab classifies user queries (computational tasks) and reserves
communication services of the IIP system in order to guarantee the QoE for the user.
The computational tasks are scheduled in order to minimize the waiting time of the
user, which is done by computational service monitoring and dynamic configuration
of communication links using the IPv6 QoS infrastructure. This approach is used to
address the requirements defined in the introductory section:
1. Content awareness – such OL-Service is chosen to provide the minimum
processing time. The data volume of the task influences the parameters used during
the link reservation in IPv6 QoS system (to achieve the minimum transfer time).
2 Context awareness is maintained by the Load Balancer. Its task is to analyze the
stream of requests and manage the negotiations with the service stratum. It is also
equipped with the prediction module which forecasts the user behavior.
3 User awareness. The services are personalized, taking into account the user
preferences, typical volumes of associated data and recommendation scheme.
Taking the above into account the general task of Online Lab is to compose a
computational service, given the request stream from the users is known or predicted.
All the components of the system (OL-Core and available OL-Services) are registered
and have unique IDs. Once the optimal (with respect to the QoE) structure of this
complex service (including the set of OL-Services and the parameters of
communication links between them) is decided by the Load Balancer, the OL-Core
reserves (via SCF functions, as described in sec. 2) the communication links connecting
all the Online Lab services. This guarantees delivery of required QoS parameters. In the
second phase the negotiation in the service stratum takes part to establish and confirm
pairwise communication between the services. After that the computational tasks are
scheduled and assigned to the appropriate OL-Services by the OL-Core.
An additional unique feature of Online Lab is the possibility of implementing
dedicated computational services which may be available to other applications. An
example of this scenario will be sketched in the following section, where we describe
the use of Online Lab service to be used by the SmartFit application.
3.3 Custom Application

The complex service model assumed for IPv6 QoS applications makes it easy to
develop intelligent applications build on the basis of available communication and
computational services. Intelligence lies in: distributed measurement data acquisition
and processing, decision making support, ability to compose services tailored to the
user’s needs and profile.
The following atomic services are proposed as building blocks for complex
services design. Computational services are: System Identifier (SI) to fit mathematical
model [5] to the user, System Simulator (SS) that makes use of the user model to
predict the user reactions for different activities, System Controller (SC) that gives
advices to the user, Training Protocol Optimizer (TPO) that works out the best
training protocol for the task the user is about to perform. Communication services
exchange data between measurement devices, services and users end-points and, if
necessary, perform additional actions, such as encryption. Using the set of atomic
services, the systems is able to deliver advanced functionalities.
The system supports the user in controlling intensity of an exercise. In this custom
application the data processing needed to support real-time training procedure is
performed by a dedicated OL-Service which serves as a decision-making component
of SmartFit application.
Typical usage scenario starts from calling Training Protocol Optimizer (TPO) that
generates optimal training protocol and passes it to the System Controller (SC) and the
System Simulator (SS). The System Controller (SC) uses Heart Rate signal, sent by the
measurement device, to support the user in maintaining exercise intensity on the desired
level. The System Simulator (SS) compares the user model response to signals obtained
from measurement devices. After an exercise is finished, measurements are sent to the
System Identifier (SI) to validate the model and update it.
Configuration of services depends on the sport discipline practiced by user (context
awareness). The system reacts to long term effects caused by systematic training by
adjusting the user model and the choice of services he/she needs (user awareness).
Note, that the above scenario is just an example of personalization scheme which
relies on composition of new services on the basis of atomic services (SI, SS, SC,
TPO). The final functionality of the resulting complex service is suited to the user
requirements. New services, originating from different applications atomic services,
may be also included and taken into account when composing complex services. For
instance, when we want to deliver the same application for a diabetic user, we may
compose a new service that supports physical training, taking blood glucose level into
account (Fig. 5), [11]. In such a case, additional service to keep blood glucose level
within normal range is executed and additional requirement to blood glucose data
encryption (content awareness) is imposed. Additional constraint for TPO is defined
to make sure that the training scenario generated by TPO will be safe for the user.
Bl
ue
to
ot
h
Fig. 5. Network topology for exemplary application
Described application provides context, content and user aware adaptive control
system supporting the user in a real-time.
4 Conclusions
In this work we presented a general idea of delivering to end-user complex services in

a distributed networking environment. The main feature of the proposed idea is that
the process of complex services delivery is aware of the content being delivered, the
context of the services delivery and that the delivered services are personalized for
each separate end-user. In order to achieve the content, context and user awareness we
proposed a general scheme for signaling system, which task is to configure distributed
application modules and network resources with respect to the requirements imposed
by the content being delivered, context of services delivery and specific user’s needs.
The proposed signaling system was designed and implemented as a middleware
between the end-user and the network, more specifically as a service stratum in the
NGN architecture of IPv6 QoS system. Note, however, that the signaling system
architecture does not assume any specific network architectures. The idea is to utilize
communication services provided by the network layer to provide fully customizable
application layer services built from computational and communication resources
available in the distributed system. This means that if the application use case scenario
does not require any specific quality of communication services it can be delivered to
the user with use of nowadays best-effort IPv4 internet network infrastructure.
The IP QoS infrastructure and all applications presented in this work are results of
the IIP project [4]. The ongoing research efforts are devoted to further integration of
service-based applications within this scheme, and the testing of new software-building
paradigms and concepts stemming from service customization and composition
approach illustrated in sec. 3.3.
Acknowledgements. The research presented in this paper was supported by the

European Union within the European Regional Development Fund program no.
POIG.01.01.02-00-045/09.
References
1. Alemdar, H., Ersoy, C.: Wireless Sensor Networks for healthcare: A Survey. Computer
Networks 54, 2688–2710 (2010)
2. Bęben, A., et al.: Architecture of content aware networks in the IIP system. Przegląd
Telekomunikacyjny, Wiadomości Telekomunikacyjne 84(8/9), 955–963 (2011) (in polish)
3. Blake, S., et al.: An architecture for differentiated services. RFC2475 (1998)
4. Burakowski, W., et al.: The Future Internet Engineering Project in Poland: Goals and
Achievements. In: Future Internet Poland Conference, Poznan, Poland (October 2011)
5. Cheng, T.M., et al.: Nonlinear Modeling and Control of Human Heart Rate Response
During Exercise With Various Work Load Intensities. IEEE Trans. on Biomedical
Engineering 55, 2499–2508 (2005)
6. Cobelli, C., et al.: Diabetes: Models, Signals, and Control. IEEE Reviews In Biomedical
Engineering 2, 54–96 (2009)
7. Mosharaf Kabir Chowdhury, N.M., Boutaba, R.: A survey of network virtualization.
Computer Networks: The International Journal of Computer and Telecommunications
Networking 54(5), 862–876 (2010)
8. Grzech, A., Rygielski, P., Świątek, P.: Translations of Service Level Agreement in
Systems Based on Service-Oriented Architectures. Cyb. and Systems 41, 610–627 (2010)
9. ITU-T Rec. Y. Functional requirements and architecture of next generation networks
(2012)
10. Jacobson, V., Smetters, D.K., Thornton, J.D.: Networking Named Content. CACM 55(1),
117–124 (2012)
11. Man, C.D., et al.: Physical activity into the meal glucose-insulin model of type 1 diabetes:
in silico studies. Journal of Diabetes Science and Technology 3(1), 56–67 (2009)
12. Noguez, J., Sucar, L.E.: A Semi-open Learning Environment for Virtual Laboratories. In:
Gelbukh, A., de Albornoz, Á., Terashima-Marín, H. (eds.) MICAI 2005. LNCS (LNAI),
13. Pautasso, C., Bausch, W., Alonso, G.: Autonomic Computing for Virtual Laboratories. In:
Kohlas, J., Meyer, B., Schiper, A. (eds.) Dependable Systems: Software, Computing,
Networks. LNCS, vol. 4028, pp. 211–230. Springer, Heidelberg (2006)
14. Świątek, J., Brzostowski, K., Tomczak, J.: Computer aided physician interview for remote
control system of diabetes therapy. In: Advances in Analysis and Decision-Making for
Complex and Uncertain Systems, Baden-Baden, Germany, August 1-5, vol. 1, pp. 8–13
(2011)
15. Tarasiuk, H., et al.: Provision of End-to-End QoS in Heterogeneous Multi-Domain
Networks. Annals of Telecommunications 63(11) (2008)
16. Tarasiuk, H., et al.: Performance Evaluation of Signaling in the IP QoS System. Journal of
Telecommunications and Information Technology 3, 12–20 (2011)
Towards a Narrative-Aware Design Framework
for Smart Urban Environments
Lara Srivastava1,* and Athena Vakali2,**

1
Department of Media Communications, Webster University, Geneva, Switzerland
[email protected]
2
Department of Informatics, Aristotle University of Thessaloniki, Greece
[email protected]
Abstract. Innovation in smart city systems is based on the principle that

devices, places and everyday things can each be enabled to serve people in a
real-time and responsive manner. This chapter presents a novel approach to the
design of smart city systems that takes into account not only technical
installations in a future Internet of Things environment, but also the power of
human storytelling in an always-on networked world. It is only when
environments are both sensor-driven and socially-aware that a more holistic,
and therefore more useful, urban narrative can emerge in the future Internet
context. The present chapter proposes a new narrative-aware design framework
and applies it to a hypothetical city scenario in order to highlight its main
components and the benefits it may offer to a future Internet city's actors.
Keywords: Smart cities, sensor data analysis, social data mining, smart urban
services, Internet of things, narrative, storytelling, navigation, mobility, sensors,
web 2.0.
1 Introduction
The Internet of today enables users to access an unprecedented amount of information
at anytime and from any device. In the future, an emerging Internet of Things (IoT)
will connect everyday objects (such as toothbrushes, shoes or car keys), which will
become information storehouses of their own, capable of collecting and transmitting
real-time data to their surrounding environment (people, places and things). The
resulting myriad of smart interconnected objects and places will make up the
intelligent urban landscape of the future.
Urban environments offer unique opportunities for developing and testing new
applications and platforms in line with the vision of the Internet of Things. European
IoT platforms have already begun emerging over the last few years inline with the
*
Professor.
**
Associate Professor.
Towards a Narrative-Aware Design Framework for Smart Urban Environments 167
future Internet momentum. Large smart city infrastructures have now been set up in
Europe (e.g.SmartSantander1, Spain) and worldwide (e.g. Songdo2, South Korea).
The growing need and interest in smart city innovation was highlighted by the
Commission in its report “Internet of Things in 2020: A roadmap for the future”3, in
which it identified key topics such as “Smart living” as part of what it termed a
“mastered continuum of people, computers and things”. There is a growing number of
innovative social and people-centric application areas, including social networking,
smart metering, smart data collection, city information models and so on [Atzoria10].
Although these application areas provide an excellent starting point to test services
and infrastructure, most offer merely quantitative solutions for a world that is
primarily qualitative (particularly from the human perspective). For the most part,
they collect and store data and information from technical devices and sensors. With
the growth of web 2.0 and social media, however, a wide array of human experience,
information and know-how is being shared and distributed across networks –
information that has yet to be properly harvested for the creation of smarter living
environments.
The present chapter proposes a new design framework for the smart city, one which
considers quantitative sensor-generated data as well as qualitative human-generated data
through participatory web platforms, in the future Internet context. In this manner,
storytelling and “listening” by networked objects is enhanced and vetted by human
storytelling, thereby getting us that much closer to true human-machine collaboration.
This chapter begins with an overview and gap analysis of the main developments
in urban IoT applications with a focus on resident mobility (Section 2). It then goes
on to highlight the need for a new kind of holistic urban storytelling (Section 3). The
section that follows describes a new design approach for smart urban environments
that is both sensor-driven and socially aware (Section 4). The concept is then applied
to a hypothetical urban mobility scenario (Section 5).
2 Urban Mobility: State of the Art and Gap Analysis

Smart city platforms and installations continue to expand as IoT innovations emerge
and services develop. At the same time, urban residents continue to multiply the
number of mobile devices they use, and through social media, have become important
generators of content themselves (with varying levels of objectivity). As mentioned
above, applications can benefit from the combination of smart city sensors and data
from users on the move.
Mobility services can capture the pulse and momentum of a city, through sensors,
status updates, and tracking. Thus far, urban mobility solutions have relied primarily
on information and communication technology to manage transport networks and
1
http://www.smartsantander.eu/
2
http://www.songdo.com/
3
ftp://ftp.cordis.europa.eu/pub/fp7/ict/docs/enet/internet-of-
things-in-2020-ec-eposs-workshop-report-2008-v3_en.pdf
168 L. Srivastava and A. Vakali
Table 1. Urban mobility: Current and future requirements
Area Current developments Future Needs and Requirements

• Lower hazardous emis- monitor city traffic based on
sensor-aware sions from city traffic multiple inputs (sensors and
transport • Environmental, accurate social streams)
management and cost-efficient road  leverage real time feedback
traffic management from residents on the move)
• plan navigations from A to  increase personalized urban
urban B by using particular travel planning aware of both
travel location and transport environmental and residents
planners modes. status
• Generic planning provide safe residents’ mobility
recommendations on pre- by data integration
defined fixed city’s
touristic spots.
• social networking leverage Web 2.0 capabilities to
social-wise supports urban mobility allow residents expressing
urban mobility primarily by POI and their city sensing
guidance user’s ratings exploit active and real-time
• location based social sensor /social capabilities
networks offline analysis for urban notifications and
alerting
guide users through a city. Examples include European projects such as SMILE 4
which deals with sustainable and cost-efficient mobility and i-Travel5 which provides
personalized, context-aware “virtual travel assistant” services in urban settings.
The majority of existing ICT urban mobility applications have focused on :
• Sensor-aware transport: This area deals with effective traffic management for a
city’s public transportation system. Sensors (e.g. in combination with IoT
platforms) capture specific measurements (such as CO2 emissions) [6]. Use cases
have focused on managing city traffic, eco-driving and emergency handling [9].
• Urban travel planners: Planners are generated on the basis of the current location
of users, their preferences and mobile device settings. Semantic web tools and
technologies such as Global Navigation Satellite Systems (GNSS) and
Geographical Information Systems (GIS) are used to improve context and geo-
location awareness, respectively [12]. Current mobile route planning tools are
typically geared towards points of interest for tourists (sightseeing, hotels,
restaurants, and packaged tour routes) [14], [17].
4
SMILE : Towards Sustainable Mobility for people in urban areas http://www.ist-
world.org/ProjectDetails.aspx?ProjectId=258180ce08fd44cfa050fc
554c80e828
5
i-Travel : The connected traveler in the city, region and world of tomorrow http://
cordis.europa.eu/fetch?CALLER=FP7_PROJ_EN&ACTION=D&RCN=85751
• Social-wise urban mobility guidance: With the emergence of web 2.0, some urban
mobility applications have sought to leverage the opinions of users on urban points
of interest (POIs). Collaborative filtering approaches are combined with location-
based partitioning and user comments [15]. More recently, recommendations for
POIs have used location-based social networks (e.g. Foursquare) and included user
ratings, proximity and similarities [16], [13]. Services such as GeoLife analyze the
GPS trajectories of users off-line to provide personalized recommendations [18].
Table 1 summarizes current practices in the three areas identified above. It identifies
some of the more pressing needs and priorities for each area. It would seem evident
that little effort has been made to date to exploit the synergies between technical data
and social data streams.
It is important to note that both the IoT industry and the mobile industry are
continuing to expand, though not always in the same direction. Therefore, much is to
be gained by unifying their vision and creating a more holistic understanding of user
needs and requirements. The convergence of machine and human perspectives will
serve to enrich and facilitate daily living in the urban context. The next sections
propose a new design for urban mobility based on this principle.
The needs and requirements (summarized in Table 1) reveal that a proposed
approach which will merge real-time social and sensor data streams is expected to be
beneficial since citizens engagement can be improved. Such an improvement is
guaranteed by the fact that, according to the authors knowledge, there are no universal
applications which go beyond a typical residents navigation or mobility assistance.
3 Towards a Richer Form of Urban Narrative
In an Internet of Things environment, individual objects collect data through a

combination of sensor and location tagging. The way these objects then transmit
information about the world around them helps people begin to create an urban
narrative that spans both space and time. In other words, these bits and bytes
contribute to the “digital memory” of a place, a city, or an event. Such digital
memories are facilitated in the future Internet context.
However, the information revealed by sensors provides only half the picture. It is
limited to data that devices have gathered through technical means- the sounds they
hear, the temperature they detect. But if sensor data is then combined with the many
layers of human observation and perception (human storytelling), what results is a
richer, more holistic “urban digital narrative”. For not only will sensors report back
with real-time recorded data, but these reports will be further checked and enhanced
by the human perspective. This will have the effect of expanding the role of sensors
beyond mere data-listeners. By way of analogy, sensors will no longer only hear the
“words”, but they will also be able to “listen” to the stories; they will not only collect
fragments of data, but be part and parcel of a larger human picture.
3.1 Sensor Data

Today's urban Internet of Things infrastructure consists mainly of automatic
identification and sensor installations connected with local, global, public and private
networks.
These networked sensors are widely used in the context of smart cities in order to
accurately measure and record physical parameters and phenomena that have an
incidence on the life of residents. They automatically listen to the environment in
which they are placed according to how they are programmed. With the right
analytical methods, this data can be used to identify and predict when special
conditions or phenomena arise, e.g. traffic congestion or atmospheric pollution. In
turn, this information can assist authorities in taking short-term (e.g inform/alert
residents and especially vulnerable social groups) or long-term action (e.g.
understanding the progress of the phenomena and trying to address them in order to
improve the city’s conditions).
Even smart sensor networks that capture information about the technical
parameters of a given phenomenon cannot share or combine this information with the
perception or mood of residents, and are thus far from creating a complete urban
digital story. In the future Internet reality, there is a need for completing existing
smart systems with people-engaged narratives that are not only technical but social in
nature, where the objective is not only to listen to the “data”, but to support the
exchange (story-listening and story-telling) of urban “stories”.
3.2 Social Stories
The daily social activities of residents are being broadcast in real-time by a growing
number of mobile devices. Urban residents use mobile devices to manage their
professional and personal lives, their interaction with others, and their interaction with
their environment. Not surprisingly, the use of web 2.0 applications and social media
has proliferated on mobile networks. In this context, mobile users act as storytellers
and listeners, exchanging experiences over the internet. This so-called “urban social
pulse” can be gleaned through applications like Flickr, Facebook, Twitter but also
location-based services such as Foursquare and Gowalla. It contributes to a larger
urban story that can be heard by authorities and residents. For instance, residents
might express overcrowding and excessive heat (e.g. at a concert), and this might
serve to override physical data such as room temperature (which may not be high
enough to cause concern). Future concerts planned in that area could be re-
considered. In this manner, a future internet would augment sensor-generated data
into usable stories that might refine a resident-driven urban narrative.
3.3 From Data Listening to Story Telling
Figure 1 is a graphical representation of the two sides of an urban story (sensor data
and resident data) and how these fit within the smart city context.
Fig. 1. Urban data listening and story telling
As illustrated, the storytelling of residents on the move can occur in real-time

through the use of mobile devices and web 2.0 platforms. Simultaneously, and also in
real-time, sensors listen to things and detect environmental conditions (in various
places). A clearer and richer urban digital narrative emerges only at the intersection of
these two data streams. Smart urban applications that can listen to this more complete
narrative are likely to provide richer, more people-centric and context-aware services.
Without a doubt, the time is ripe to take urban mobility services this important step
further. The notion of context-awareness in the future internet need not be limited to
sensory data or machine-generated information, but must also leverage the power of
human storytelling.
3.4 Uses and Benefits
In the future internet, sensor and socially-aware storytelling could provide vital
support and guidance to a city’s actors, such as residents, visitors and authorities.
Uses and benefits for these actors are highlighted below:
• Residents typically move around their city on a daily basis using different modes of
transport (e.g. personal vehicles, public transport, walking, running, biking) and for
various purposes at different times of the day (e.g. work, caregiving, errands,
leisure etc.). Narrative-aware services should place emphasis on the collection of
information that generates real-time adaptive recommendations for residents. This
can improve navigation within the city and can assist in the selection of the most
appropriate routes based on various parameters, e.g. distance, CO2 emissions,
congestion, noise levels, parking, public transport routes and schedules. Resident
input (stories) over social media can provide invaluable qualitative information to
complement sensor-generated data.
• Visitors are particularly interested in city navigation, POIs, queues and crowds.
Narrative-aware services could offer recommendations on points of interest or city
walks, based on proximity, popularity, tourist opinions, weather, opening times,
congestion and so on. Both residents and visitors could provide observations that
would be used to complement any technical measurements taken by sensors and
location-based technologies.
• Authorities (such as the police, fire department, or city council) could exploit
narrative-aware urban services and applications to enable monitoring of the city’s
major “variables” (e.g. noise, temperature, crime, CO2 levels) through global and
user-centered visualization interfaces. Such interfaces could also enable detection
of vulnerable geographical areas which indicate both over-threshold sensor
measurements and any alerts broadcast by the residents on the move.
The fact that the proposed narrative approach is flexible and multi-scenario oriented,
differentiates it from existing approaches which are more vertical and they focus on
improving the separate angles, i.e. either the data management or the usage side. The
end-users involvement is expected from the appealing story telling emphasis which is
expected to attract in particular users with mobile phones in a smart city context.
4 The Urban Narrative-Aware Design Framework
Capturing and reading urban narratives involves several complex steps and processes,
and cuts across various service layers (infrastructure applications, content, usage).
The design of such a framework must be cognizant of this complexity. Figure 2 is an
illustration of the urban narrative-aware design framework, with functionality at three
different levels:
• Data and stories: All tasks related to the collection of data and socially-generated
stories are carried out at this level. Differentiated techniques are required for
storing sensor data and social streams into individual “DataStores” or data
repositories. Targeted data storage and scalable indexing schemes should be used
to cope with the ever-growing number of sensor and social measurements.
Moreover, specific data and stories pre-processing is required to provide noise-free
DataStores.
• Analysis and Processing. DataStore integration, refinement and analysis are key
tasks at this level. In particular, the first core task is the integration of sensor data
and human stories, with the objective of constructing new narrative-aware
DataStores, i.e. “Narrative stores” or “NarraStores”. NarraStores host the various
digital narratives of an urban context. This integration is an ongoing task which can
benefit from regular refinement (i.e. calibration) and analysis, due to the emerging
and unpredictable nature of urban sensor and social data streams. DataStore
calibration involves processes which will validate and fine-tune information from
the two different data sources (sensors and social) and will revise either the content
of the DataStore or the data collection process itself. DataStore analysis can
involve a wide array of methodologies and algorithms from the fields of data
mining and recommender systems. This leads to the generation of quantitative and
qualitative NarraStores. The continuous calibration and analysis processes
optimize the content of NarraStores by summarizing, clustering and packaging
diverse narratives.
• Services and applications. NarraStores are the foundation upon which a wide
variety of services and applications can be built. The proposed design in Fig.2
puts forward an initial number of key urban services (assistance, alerting, and
planning) for the main urban actors (as outlined in 3.4). Such services leverage
the Narrative stores at their disposal in order to offer contextualized mobile and
Web applications for short-term (e.g. alerting) or long-term tasks (e.g. business
planning).
Fig. 2. A smart urban narrative aware design framework
5 Challenging the Design: An Urban Mobility Scenario
The narrative-aware design framework proposed herein is naturally tailored to the

urban context, which is characterized by the need for flexible, scalable and adaptive
processes in future Internet settings.
This section describes examples of urban applications based on the design
framework, and then applies these examples to a specific event, namely a jazz
festival. It is assumed that the urban context is one of a highly populated city with
smart sensor installations at various locations (parking places, city lights, transport
buses etc). The fictitious city in this case is known as “SmartVille”. SmartVille would
also have in place a dedicated City Department, whose objective would be to support
narrative-aware services under the proposed design framework.
5.1 Examples of Narrative-Aware Urban Applications
SmartVille hosts and supports scalable DataStores and the resulting NarraStores,
which are repositories stored in large scale data centers. As described in the previous
section, these NarraStores include both SmartVille's physical data (generated by
sensors and IoT installations) and its social data streams (generated by residents on
the move). The NarraStores collect social data in an anonymous manner, such that
no personal or private data is traceable. These NarraStores are then offered by city
authorities either for public sector use or for private enterprises, for the development
urban applications and services.
Table 2. An indicative list of smart narrative-aware services
Narrative stores Processes

Service Description Integration Analysis Calibration
City- Monitors parking slots and Real-time 
Park recommends alternatives;
Offered by the City
City- Emergency alerting Real-time  
Watch (weather, fire, etc);
Offered by the City
City- City Navigation opportu- Asynchronous  
Nav nities in a eco-friendly
manner; Offered by
mobile vendors
City- Capture residents opinions Asynchronous  
Pulse and moods; Offered by
mobile vendor to the
authorities
The city's NarraStores are made available to these public and private clients
through a cloud infrastructure, as follows:
• Resident cloud services: Resident-specific information regarding daily urban
living, public places and things with common safety concerns (in coordination with
city police and other related departments). Costs are sponsored by the City and
services are therefore offered to residents on a discounted basis.
• Emergency cloud services: Emergency information and services available to all
residents, visitors and the wider general public over the Internet and mobile
networks (free of charge)
• Public cloud services: Information, navigation and emergency services available to

residents, visitors and the wider general public over mobile networks and the
Internet (free or offered on a pay-per-usage model).
• Private cloud services: offered to private organizations (e.g. mobile or Internet
vendors) to encourage the development of advanced and innovative city services
(at market cost).
A number of services and applications in Smartville use the cloud infrastructure. Four
examples are set out in Table 2 below. Technology developers and providers harvest
the City's NarraStores to create new services and applications. They use the narrative-
aware design processes (e.g. analysis, calibration) to constantly improve and enhance
their offerings. Table 2 sets out the type of integration (real-time or asynchronous),
and the amount of analysis and calibration required for the service in question.
For example, the SmartVille City-Watch service uses the integrated NarraStores,
such that emergencies are reported in realtime and in context, through both physical
and social data. Such emergencies are captured by peak activities recorders at the
sensor end (e.g. excessive heat) verified by simultaneous social bursts (e.g. heavy
tagging on fire) through Web 2.0. Careful analysis (tag-based clustering) and
algorithms (collaborative filtering), are applied to the DataStores in order to alert
users of emergencies.
Another example is the SmartVille City-Pulse service, which monitors NarraStore
content by continuously analyzing and calibrating DataStores content asynchronously
(i.e. not necessarily in real-time). It uses appropriate time-aware clustering on the
DataStores (to reveal sensor/social groups) to make cluster-based recommendations
(e.g. in the form of tag topic clouds). As a result, City-Pulse reports the most
prominent topics highly ranked by residents during a particular time period.
5.2 Narrative-Aware Applications in Action: The Jazz Festival
Events like festivals and concerts are quite popular in SmartVille. For instance, the
city hosts an annual 3-day jazz festival, attracting many residents and visitors who use
the services highlighted above: City-Watch, City-Park, City-Nav and City-Pulse.
Table 3. Use of narrative aware services in practice
City-Park City-Watch City- City-Pulse

Nav
pre-event arrangements    
at-event assistance  
post-event facilitating  
future-event planning    
During the jazz festival, there are scheduled shows and concerts on various stages
around the city, e.g. in the SmartVille stadium at two major city squares. Locations
are equipped with IoT sensor installations and are not far from main traffic routes.
In this context, the proposed narrative-aware services can support the jazz festival in all
of its phases and create a more pleasant and engaging experience for the festival goer.
The need for each service at different phases of event planning is highlighted by
the number of checkmarks in Table 3. The analysis is based on the following:
• City-Park : recommends parking spots on the basis of relevant data tracked by
sensors; recommendations are delivered to festival attendants arriving or departing
from festival sites, on the basis of location and tagging reports (captured by their
mobile LBSN);
• City-Watch issues alerts when both sensors and social bursts report emergency
situations within the festival environs; these alerts are of importance prior, during
and after the festival, since they can monitor the entire span from sensor to social
threads (e.g. from temperature levels to overcrowding, respectively);
• City-Nav delivers navigation recommendations (primarily) to visitors for an eco-
aware and safe arrival/departure at the festival sites;
• City-pulse stores and monitors narratives during the entire festival through the
fusion of sensors and social stories. As a result of this continuous processing, the
narratives offer valuable information to authorities. There is a triage and ranking of
narratives such that important conclusions can be reached after, during and after
the event. For example, authorities might reorganize parking areas and re-program
sensor installations in line with user demand for parking. It might also be possible
to verify whether certain sensors are malfunctioning (e.g. sensor oversensitivity),
as sensor data can be refined or even contradicted by social storytelling.
The design framework remains abstract in its main design principles in order to offer
a wide range of potential uses and scenarios. Its simplicity can supports different
applications and services which might range from the event managing and scheduling
to new policy making.
6 Conclusion
Urban environments offer a fertile ground for developing and testing new smart
applications in line with the Internet of Things and the future Internet vision. The
narrative-aware design framework proposed herein exploits sensor and social data
collection in a holistic manner through its design integration, analysis and calibration
processes. The design includes qualitative data stores (and not merely quantitative ones)
which embed both machine (sensors) and human (social) measurements. Alerting,
assistance and planning are considered vital services in a city context, as highlighted in
the event-based scenario. Narrative-aware design can be of tremendous benefit to
primary future Internet city actors (residents, visitors and authorities) for a wide range
of services and requirements (e.g. time-critical, long-term analysis, processing rates etc).
Such a holistic approach is invaluable for the development of the smart, context-aware
and user-centric services that lie at the very heart of a future Internet.
References
1. Atzoria, L., Ierab, A., Morabitoc, G.: The Internet of Things: A survey. Computer
Networks 54(15), 2787–2805 (2010)
2. Breunig, M., Baer, W.: Database Support for Mobile Route Planning Systems. Computers,
Environment and Urban Systems 28(6), 595–610 (2004)
3. Castillo, L., et al.: SAMAP: An User-Oriented Adaptive System for Planning Tourist
Visits. Expert Systems with Applications 34(2), 1318–1332 (2008)
4. Chiu, D.K.W., Leung, H.F.: Towards Ubiquitous Tourist Service Coordination and
Integration: a Multi-Agent and Semantic Web Approach. In: International Conference on
Electronic Commerce, pp. 574–581 (2005)
5. Driver, C., Clarke, S.: An Application Framework for Mobile, Context-Aware Trails.
Pervasive and Mobile Computing 4(5), 719–736 (2008)
6. Giannopoulos, G.A.: The application of information and communication technologies in
transport. European Journal of Operational Research 152 (2004); [Guinard10] Guinard, D.,
Trifa, V., Wilde, E.: A resource oriented architecture for the Web of Things. In: Internet of
Things (IOT) Conference (December 2010)
7. Guinard, D., Trifa, V., et al.: Interacting with the SOA-Based Internet of Things:
Discovery, Query, Selection, and On-Demand Provisioning of Web Services. IEEE
Transactions on Services Computing 3(3), 223–235 (2010)
8. Kenteris, M., Gavalas, D., Economou, D.: An Innovative Electronic Tourist Guide
Application. Pervasive and Ubiquitous Computing 13(2), 103–118 (2009)
9. Hernández-Muñoz, J.M., Bernat Vercher, J., Muñoz, L., Galache, J.A., Presser, M.,
Hernández Gómez, L.A., Pettersson, J.: Smart Cities at the Forefront of the Future
Internet. In: Domingue, J., et al. (eds.) FIA 2011. LNCS, vol. 6656, pp. 447–462. Springer,
Heidelberg (2011)
10. Niaraki, A.B., Kim, K.: Ontology Based Personalized Route Planning System Using a
Multi-Criteria Decision Making Approach. Expert Systems with Applications 36(2p1),
2250–2259 (2009)
11. Prete, L.D., Capra, L.: diffeRS: A Mobile Recommender Service. In: Proceedings of the
2010 Eleventh International Conference on Mobile Data Management (MDM 2010), pp.
21–26. IEEE Computer Society, Washington (2010)
12. Sadoun, B., Al-Bayari, O.: Location Based Services Using Geographical Information
Systems. Computer Communications 30(16), 3154–3160 (2007)
13. Sandholm, T., Ung, H.: Real-time Location-aware Collaborative Filtering of Web Content.
In: Proceedings of the 2011 Workshop on Context-awareness in Retrieval and
Recommendation (CaRR 2011), pp. 14–18. ACM, New York (2011)
14. Scherp, A., Boll, S.: Generic Support for Personalized Mobile Multimedia Tourist
Applications. In: ACM International Conference on Multimedia, pp. 178–179 (2004)
15. Takeuchi, Y., Sugimoto, M.: CityVoyager: An Outdoor Recommendation System Based
on User Location History. In: Ma, J., Jin, H., Yang, L.T., Tsai, J.J.-P. (eds.) UIC 2006.
16. Ye, M., Yin, P., Lee, W.-C.: Location recommendation for location-based social networks.
In: Proceedings of the 18th SIGSPATIAL International Conference on Advances in
Geographic Information Systems (GIS 2010), pp. 458–461. ACM, New York (2010)
17. Yu, C.C., Chang, H.P.: Personalized Location-Based Recommendation Services for Tour
Planning in Mobile Tourism Applications. In: Di Noia, T., Buccafurri, F. (eds.) EC-Web
2009. LNCS, vol. 5692, pp. 38–49. Springer, Heidelberg (2009)
18. Zheng, Y., Zhang, L., Xie, X.: Recommending friends and locations based on individual
location history. ACM Trans. Web 5(1) (2011)
Urban Planning and Smart Cities:
Interrelations and Reciprocities
Leonidas G. Anthopoulos1,* and Athena Vakali2,**

1
Project Management Department, TEI of Larissa, Greece
[email protected]
2
Computer Science Department, Aristotle University of Thessaloniki, Greece
[email protected]
Abstract. Smart cities are emerging fast and they introduce new practices and
services which highly impact policy making and planning, while they co-exist
with urban facilities. It is now needed to understand the smart city’s contribution
in the overall urban planning and vice versa, to recognize urban planning
offerings to a smart city context. This chapter highlights and measures smart city
and urban planning interrelation and identifies the meeting points among them.
Urban planning dimensions are drawn from the European Regional Cohesion
Policy and they are associated with smart city’s architecture layers.
Keywords: Smart city, digital city, sustainability, urban planning, regional

planning.
1 Introduction
Regional planning concerns the context and the organization of human activities in a
determined space via taking into account the available natural resources and the
financial requirements. Urban planning particularizes regional planning in a
residential area. Both regional and urban planning are policy frameworks that reflect
the Government willing for sustainable land uses and development in a specific space
for a limited time period [6], [9], [12], [14]. Planning accounts various parameters
such as the environmental capacity, population, financial cohesion, and transportation
and other public service networks.
Smart cities appeared in late 80s as a means to visualize urban context and they
evolve fast since then. Today, they enhance digital content and services in urban
areas, they incorporate pervasive computing and they face environmental challenges.
Various international cases present alternative approaches to the smart city, while
they capitalize the Information and Communication Technologies (ICT) for multiple
purposes, which vary from simple e-service delivery to sophisticated data collection
for municipal decision making. South Korean smart cities for instance, use pervasive
*
Assistant Professor.
**
Associate Professor.
Urban Planning and Smart Cities: Interrelations and Reciprocities 179
computing to measure various environmental indices [15], which are used by the local
Government to carry out interventions for the improvement of life in the city (e.g. for
traffic improvement).
This chapter is inspired by the co-existence of the smart city and the urban space,
and seeks to investigate the relation between the smart city and the urban planning, in
terms of mutual support and benefit. In order for this relation to be identified, an
analysis of these terms and of their structure is performed, and the points of mutual
interest are recognized. Moreover, this chapter addresses the Future Internet
application areas that comprise out of user areas and communities, where the Future
Internet can boost their innovation capabilities. In this context, various smart city’s
infrastructure and applications can contribute to urban planning data collection and
decision making by the planning stakeholders’ groups.
In the following background section the notions of regional and urban planning are
described and the planning framework is outlined on the basis of the European
practice. Moreover, the smart city context is clarified, along with a classification of
various metropolitan ICT-based environments which are further evaluated according
to a generic architecture. Section 3 identifies and summarizes interrelations between
urban planning and smart city contexts. The final section 4 has the conclusions of this
chapter and some future implications.
2 Urban Planning: Principles and Dimensions
Various relations configure an urban space, such as financial, environmental and

social [14], which extend the notion of a city beyond a simple land formulation.
Urbanism exist for more than 5,000 years and cities were formed according to
variants such as the physical topography, the distance from and the position of the
sea, the ordinance of rivers and the transportation networks that connect cities. Forms
such as disorder, radius planning, Hippodamus planning and metropolis are the most
usual [14]. In the mid-19th century the urban and the regional planning arose as a
reaction against the industrial cities, in order to provide with some rules for
environmental and for cultural protection, and to determine future national
development.
Legislation authorizes the State to control planning’s implementation and it defines
the dimensions of the regional and the urban planning (depicted in Fig. 1) [1], [7].
These dimensions meet built environment dimensions [9] and they refer to the
following:
• Environmental protection (Quality): it deals with qualitative criteria such as:

livability, environmental quality, quality of life [11] and respect on biodiversity. In
this context planning delimits the urbanization zones, the seashore and streams;
• Sustainable residential development (Viability Timeline): it covers the urban
viability timeline since it “meets the needs of current generations without
compromising the ability of future generations to meet their needs and aspirations”
[11];
180 L.G. Anthopoulos and A. Vakali
• Resources’ capitalization (Capacity): it concerns both natural and human

resources’ capitalization with means of optimal demographic allocation and
decentralization, water and other natural resources’ use, residential and farming
allocation etc;
• Coherent regional growth support (History and Landscape): it embraces the urban
history and landscape and it is based on various Government programs’ planning
and implementation, which respect traditional settlements, archaeological areas,
forests and parks.
Fig. 1 outlines the dimensions and the hierarchical organization of a representative

European regional planning set of frameworks [5], which follows the European
directives for sustainable land use and development. According to this suggestion,
planning’s dimensions are allocated to particular frameworks: (a) the general
framework for long-term (15 years) national sustainable development; (b) the
regional framework that focuses on peripheral long-term development; (c) the special
frameworks that concern specific productivity sectors. Each particular framework
contains studies and drawings that determine:
• Demographic distribution that concerns the Capacity dimension;

• Land uses that meet the Quality and the History and Landscape dimensions;
• Transportation and other utility infrastructures that align to Capacity dimension;
• Forests and parks that concern both the Quality and the Viability Timeline
dimensions;
• The environmental protection framework that contributes to the Quality dimension;
• The authorities that monitor and evaluate the planning rules that meet all of the
framework’s dimensions.
In this context, the regional planning [5], [11] seeks to protect the environment and to
secure the natural and cultural resources, while it highlights the competitive
advantages of different areas. Moreover, it strengthens the continuous and balanced
national development via taking into account the broader supranational surroundings.
Finally, it focuses on financial and on social national cohesion via signalizing
particular geographic areas with lower growth rates.
As highlighted in Fig. 1, urban planning particularizes the regional planning in
cities and residential areas, it is composed and managed by the local Governments [5],
and it is realized via three core plans (Fig. 1):
• The master plan for the metropolis.

• The general urban plan for the residential and for the suburban organization of the
cities and towns. It consists of various studies such as the urban study, the
implementation act, the rehabilitation studies etc.
• The space and residential organization plan for rural areas.
Urban planning controls the development and the organization of a city, by

determining the urbanization zones and the land uses, the location of various public
networks and communal spaces, the anticipation of the residential areas and the rules
for building constructions, and of the authorization of the monitoring and of the
intervention procedures. Campbell [6] described the triangle of conflicts (property,
development and resource) that exist between economic development, environmental
protection, equity, and social justice, and which the urban planning aim to
manipulate.
Fig. 1. The hierarchical organization diagram of regional and urban planning’s framework
3 Smart Cities: Key Attributes and Characteristics
According to [8] the term smart city is not used in a holistic way describing a city with
certain attributes, but is used for various aspects which range from smart city as an IT-
district to a smart city regarding the education (or smartness) of its inhabitants. In this
context, the smart city is analyzed in intelligent dimension [8], [13], which concern
“smart people”, “smart environment”, “smart economy”, “smart governance”, “smart
mobility” and at a total “smart living”.
The term was originally met in Australian cases of Brisbane and Blacksbourg [4]
where the ICT supported the social participation, the close of the digital divide, and
the accessibility to public information and services. The smart city was later evolved
to (a) an urban space for business opportunities, which was followed by the network
of Malta, Dubai and Kochi (India) (www.smartcity.ae); and to (b) ubiquitous
technologies installed across the city, which are integrated into everyday objects and
activities.
The notion of smart city has been also approached as part of the broader term of
Digital City by [2], where a generic multi-tier common architecture for digital cities
was introduced, and assigned smart city to the software and services layer. This
generic architecture (Fig. 2) contains the following layers:
• User layer that concerns all e-service end-users and the stakeholders of a
smart city. This layer appears both at the top and at the bottom of the generic
architecture because it concerns both the local stakeholders –who supervise
the smart city, and design and offer e-services- and the end-users –who
“consume” the smart city’s services and participate in dialoguing and in
decision making-.
• Service layer, which incorporates all the particular e-services being offered
by the smart city.
• Infrastructure layer that contains network, information systems and other
facilities, which contribute to e-Service deployment.
• Data layer that presents all the information, which is required, produced and
collected in the smart city.
This generic architecture can describe all the different types of attributes needed to
support the smart city context, and which typically include:
• Web or Virtual Cities, i.e. the America-On-Line cities, the digital city of
Kyoto (Japan) and the digital city of Amsterdam: they concern web
environments that offer local information, chatting and meeting rooms, and
city’s virtual simulation.
• Knowledge Based Cities, i.e. the Copenhagen Base and the Craigmillar
Community Information Service (Edinburgh, Scotland): they are public
databases of common interest that are updated via crowd-sourcing, and
accompanied by the appropriate software management mechanisms for
public access.
• Broadband City/Broadband Metropolis, i.e. Seoul, Beijing, Antwerp, Geneva,
and Amsterdam: they are cities where fiber optic backbones -called
“Metropolitan Area Networks (MAN)”- are installed, and enable the
interconnection of households and of local enterprises to ultra-high speed
networks.
• Mobile or Ambient cities, i.e. New York, San Francisco installed wireless
broadband networks in the city, which were accessible (free-of-charge) by
the habitants.
• Digital Cities i.e. Hull (UK), Cape Town and Trikala (Greece) extension of
the previous resources to “mesh” metropolitan environments that
interconnect virtual and physical spaces in order to treat local challenges.
• Smart or Intelligent Cities, i.e. Brisbane and Blacksbourg (Australia), Malta,
Dubai and Kochi (India), Helsinki, Barcelona, Austin and others of smart-
cities networks (http://smart-cities.eu, http://www.smartcities.info): they are
particular approaches that encourage participation and deliberation, while
they attract investments from the private sector with cost-effective ICT
platforms. Today, smart cities evolve with mesh broadband networks that
offer e-services to the entire urban space. Various ICT vendors [10] have
implemented and offer commercial solutions for the smart cities.
Fig. 2. The multi-tier architecture of a digital city [2]
Ubiquitous Cities, i.e. New Songdo (South Korea), Manhattan Harbour (Kentucky,
USA), Masdar (Abu Dhabi) and Osaka (Japan): they arose as the implication of
broadband cost minimization, of the commercialization of complex information
systems, of the deployment of cloud services, and of the ubiquitous computing. They
offer e-services from everywhere to anyone across the city via pervasive computing
technologies.
Eco-cities, i.e. Dongtan and Tianjin (China), Masdar (Abu Dhabi): they capitalize
the ICT for sustainable growth and for environmental protection. Some indicative
applications concern the contribution of ICT sensors for environmental measurement
and for buildings’ energy capacity’s evaluation; smart grids deployment for energy
production and delivery in the city; encouragement of smart solutions for renewable
energy production.
Table 1. Measuring smart city’s sophistication
Virtual Knowl Broad Mobile smart Digital Ubiqui Eco-

Cities edge band Cities cities Cities tous Cities
bases Cities Cities
User 5 5 2 1 5 5 5 5
Infrastructur 1 1 5 5 3 3 5 3
e
Service 2 1 1 1 5 5 5 5
Data 5 5 1 1 5 5 5 5
The above smart city classification could be evaluated for its sophistication in the
following (Table 1), according to the matching of each approach to the generic multi-
tier architecture of (Fig. 2). The values of the above table are self-calculated
according to empirical findings [2], and they represent the contribution of each
architecture layer to the particular smart city approach. The rows of (Table 1)
concern the architecture layers, while the columns refer to the abovementioned smart
city approaches. The value entries are based on Likert scale (values from 1 to 5) [7]
and they reflect how important each layer is considered for each particular approach.
On the basis of this measurement:
• User layer accounts significantly in all approaches except in Broadband and

Mobile cities, where users mostly consume telecommunication services, while the
networks extend to most populated areas.
• The Infrastructure layer does not contribute in Virtual and in Knowledge Based
cities, while Smart, Digital and Eco-Cities can mostly focus on e-services that can
be deployed either via alternative infrastructure providers.
• The service layer has significant contribution to the approaches beyond the smart
city approach, while only a few services are offered in the other approaches. in
Virtual City approach the existence of various ICT infrastructure is not necessary,
while data and user layers are crucial for city virtualization.
• Finally, the Data layer is the basis for service delivery and thus contributes
significantly to all the approaches except from the Broadband and the Mobile
Cities, which offer telecommunication services.
These estimated values can support researchers and supervisors in selecting the
appropriate approach for their city [3] and to design and predict their city’s future
“character”.
4 Urban Planning and Smart City Interrelations
On the above attributes, various e-service portfolios can be offered in a modern smart
city [4]:
• E-Government services concern public complaints, administrative procedures at

local and at national level, job searches and public procurement.
• E-democracy services perform dialogue, consultation, polling and voting about
issues of common interests in the city area.
• E-Business services mainly support business installation, while they enable digital
marketplaces and tourist guides.
• E-health and tele-care services offer distant support to particular groups of citizens
such as the elderly, civilians with diseases etc.
• E-learning services offer distant learning opportunities and training material to the
habitants.
• E-Security services support public safety via amber-alert notifications, school
monitoring, natural hazard management etc.
• Environmental services contain public information about recycling, while they

support households and enterprises in waste/energy/water management. Moreover,
they deliver data to the State for monitoring and for decision making on
environmental conditions such as for microclimate, pollution, noise, traffic etc. (in
Ubiquitous and Eco-city approaches).
• Intelligent Transportation supports the improvement of the quality of life in the
city, while it offers tools for traffic monitoring, measurement and optimization.
• Communication services such as broadband connectivity, digital TV etc.
The smart city addresses the supranational planning policies - such as the European
Cohesion Policy [7] - that influence national planning policies and prioritize
transportation networks and accessibility, entrepreneurship, education and training,
and sustainable growth. These priorities affect all the four planning dimensions, while
the smart city with the intelligent transportation services, the e-business services, the
e-learning services, and the environmental services aligns to each of them
respectively. The following subsections highlight in detail this relation.
4.1 Smart City to Urban Planning Alignments

Both end-users and stakeholders of the smart city’s User layer are obliged to follow
the planning rules and to consult in cases of framework’s construction. Thus, the User
layer is influenced by all planning dimensions.
Fig. 3. The smart city’s layers align to urban planning dimensions
Moreover, the smart city’s infrastructures have to conform to planning rules and
not to charge the local environment or the local protected areas, while planning has to
uniformly develop smart cities across the regions for coherent development. In this
context, the Infrastructure layer meets all planning dimensions.
Concerning the Service layer, the environmental and the intelligent transportation
services align directly to the Quality and to the Viability Timeline planning
dimensions. Moreover, the e-Democracy services align to the Capacity dimension,
since public consultations and open dialogue can influence planning and express local
requirements; planning on the other hand aims to establish resource capitalization for
local development that meets local needs. Finally, the e-Business portfolio aligns to
the planning dimensions of Capacity and of History and Landscape, since tourist
guides demonstrate and can protect traditional settlements, archaeological areas,
forests and parks; while business installation services oblige enterprises to install in
business centers and in areas that do not influence sustainability.
Finally, the smart city’s data layer must be kept up to date with accurate planning
information, in order to deliver efficient and effective e-services to the local
community. This one way relation between smart city and urban planning is displayed
on (Fig. 3) and shows that the development of a smart city has to align to planning
dimensions.
4.2 Urban Planning Tracks to Smart City Layering
A vice versa relation exists too (Fig. 4), via which urban planning has to account the
existence of a smart city: the environmental data that is collected from ubiquitous
sensors has to contribute to Quality and to the History and Landscape dimensions, and
useful directions can be considered for land and for residential uses.
Furthermore, the smart city infrastructure layer consists of significant ICT facilities
-e.g. broadband networks, computer rooms and inductive intelligent transportation
loops-, which influence the Viability Timeline and the Capacity planning dimensions.
Fig. 4. Urban planning dimensions tracks to smart city layers

All these findings result in a bidirectional relation between planning and smart city
(Fig. 3), (Fig. 4), which shows that the smart city aligns to urban planning
dimensions, while the urban planning has to capitalize and to respect the existence of
a smart city. Furthermore, an important outcome would consider the rate of influence
between each urban planning’s dimension and each smart city’s layer. According to
the previous description, the interrelation would be measured with the meeting points
between dimensions and layers (Table 2).
The rows in (Table 2) represent the smart city architecture layers, and the columns
the urban planning dimensions. The calculated entries in table cells reflect the
meeting points that previously discussed. The Service layer for instance, meets the
four urban planning dimensions; three kinds of e-services address the Viability
Timeline dimension, meaning three meeting points (the value of 3) for this cell etc.
The Users layer meets all urban planning dimensions, since stakeholders can
participate in planning, while planning affects stakeholders. The Infrastructure layer
concerns resources and therefore Capacity in Urban Planning, while the Data layer
(e.g. environmental data collection via ubiquitous sensors) contributes and must be
accounted by the Quality and by the Viability Timeline planning dimensions. On the
other hand, the Viability Timeline and the Quality dimensions are mostly affected by
the existence of a smart city.
Table 2. Measuring the interrelation between planning dimensions and smart city’s layers
HISTORY & VIABILITY

QUALITY CAPACITY
LANDSCAPE TIMELINE
User 1 1 1 1
Infrastructure 1 1 1 1
Service 3 1 1 3
Data 1 1 1 1
5 Conclusions and Future Outlook
Smart cities are “booming” and various important cases can be faced worldwide,
which can be classified in various approaches and can be evaluated according to their
sophistication. All alternative approaches deliver emerging types of services to the
local communities with the use of physical and of virtual resources. This chapter
considered this co-existence of the smart city and the Urban Space and in this context
it investigated the interrelation between smart city and urban planning.
Urban planning supports sustainable local growth, it consists of four dimensions
that were recognized according to the European Regional Policy Framework, and
their context was described. A smart city on the other hand can follow a multi-tier
architecture, which can be considered generic for all particular approaches. The
analysis of the planning’s dimensions and of the smart city’s architecture layers
shows various meeting points, via which these two notions interact. More specifically,
smart city’s service layer aligns and contributes to all the urban planning’s
dimensions and various e-Services support sustainable local growth. On the other
hand, planning’s dimensions can be affected by smart city’s stakeholders via
participatory policy making, while the smart city’s infrastructure has to be recognized
and capitalized.
This chapter tried to interrelate the physical and the digital space of a smart city
with tangible measurement means in order to support Future Internet application
areas. Relative efforts have been performed in the South Korean ubiquitous cities,
where the smart city moved towards the environmental protection. This chapter’s
resulted meeting points between smart city’s layers and planning’s dimensions can
provide Future Internet research with details concerning where the developed
applications and the deployed infrastructure have to account the physical space and
the environment.
General suggestions that require further investigation concern that the smart city
has to be accounted in the regional and the urban planning frameworks, with means
that the ICT resources are capitalized for information retrieval and analysis for policy
making; while the environmental charge of a smart city has to be measured and
evaluated during regional and urban planning.
Acknowledgments. This chapter uses findings concerning the urban planning in

Europe from the dissertation developed by Tsopanis Dimitrios, which was supervised
by the authors under the MSc Program in Construction Management of the Hellenic
Open University (Greece), during the academic year 2010-2011. Moreover, this
chapter is supported by the “Enterprise Architecture for Digital Cities (EADIC)”
research project, which is being developed between 2012 and 2014, and funded by the
Greek General Secretary for Research and Development (ARCHIMEDES III
program).
References
1. Ahern, J.: Planning for an extensive open space system: linking landscape structure and
function. Landscape Urban Plann. 21, 131–145 (1991)
2. Anthopoulos, L., Tsoukalas, I.A.: The implementation model of a Digital City. The case
study of the first Digital City in Greece: e-Trikala. Journal of e-Government 2(2) (2006)
3. Anthopoulos, L., Gerogiannis, V., Fitsilis, P.: Supporting the solution selection for a
digital city with a fuzzy-based approach. In: International Conference on Knowledge
Management and Information Sharing, KMIS 2011 (2011)
4. Anthopoulos, L., Tougoutzoglou, T.: A Viability Model for Digital Cities: Economic and
Acceptability Factors. In: In Reddick, C., Aikins, S. (eds.) Web 2.0 Technologies and
Democratic Governance: Political, Policy and Management Implications. Springer,
Heidelberg (forthcoming, 2012)
5. Aravantinos, A.: Urban planning for sustainable growth (in Greek). Symmetria Publishing
(1997)
6. Campbell, S.: Green Cities, Growing Cities, Just Cities? Urban planningand the
Contradictions of Sustainable Development. Journal of the American Planning Association
(1996)
7. European Committee: Regional Policy: Bridging the prosperity gap (2012),
http://europa.eu/pol/reg/index_en.htm
8. Giffinger, R., Fertner, C., Kramar, H.: Meijers, e. & Pichler-Milanovic, N.: smart cities:
Ranking of European medium-sized cities (2007)
9. Handy, S.L., Boarnet, M.G., Ewing, R., Killingsworth, R.E.: How the Built Environment
Affects Physical Activity: Views from Urban Planning. American Journal of Preventive
Medicine 23(2S) (2002)
10. IBM: How Smart is your city? Helping cities measure progress (2009),
http://www-935.ibm.com/services/us/gbs/bus/html/ibv-smarter-
cities-assessment.html
11. van Kampb, I., Leidelmeijera, K., Marsmana, G.: Urban environmental quality and human
well-being Towards a conceptual framework and demarcation of concepts; a literature
study. Landscape and Urban Planning 65, 5–18 (2003)
12. Kiernan, M.J.: Ideology, politics, and planning: reflections on the theory and practice of
urban planning. Environment and Planning B: Planning and Design 10(1), 71–87 (1983)
13. Komninos, N.: Intelligent Cities: Innovation, Knowledge Systems and Digital Spaces, 1st
edn. Routledge, London (2002)
14. Koutsopoulos, K., Siolas, A.: Urban Geography – The European City (in Greek). Greek
National Technical University publishing (1998)
15. Lee, S., Oh, K., Jung, S.: The Carrying Capacity Assessment Framework for Ubiquitous-
Ecological Cities in Korea. In: 2nd International Urban Design Conference, Gold Coast
2009, Australia (2009)
16. Trochim William, M.: Likert Scaling. Research Methods Knowledge Base, 2nd edn
(2006), http://www.socialresearchmethods.net/kb/scallik.php
The Safety Transformation in the Future Internet
Domain
Roberto Gimenez1, Diego Fuentes1, Emilio Martin1, Diego Gimenez2, Judith Pertejo2,
Sofia Tsekeridou3, Roberto Gavazzi4, Mario Carabaño5, and Sofia Virgos5
1
HI-Iberia Ingeniería y Proyectos
{rgimenez,dfuentes,emartin}@hi-iberia.es
2
ISDEFE
{dgimenez,jpertejo}@isdefe.es
3
Athens Information Technology - AIT
[email protected]
4
Telecom Italia
[email protected]
5
Everis
{mario.carabano.mari,sofia.virgos.casal}@everis.com
Abstract. Public Safety is nowadays a priority, cornerstone and major concern

for governments, majors and policy makers in current (and future) smart cities.
Notwithstanding the foregoing, large advances in ICT technologies are foretold
to revolutionize our society and enhance our feeling of safety (and hopefully,
wellbeing). This chapter presents an introduction to three of the most promising
technological pillars considered to be spearheads in this transformation: Internet
of things, understood as the data capillarity through billions of sensors,
Intelligent Video Analytics and Data Mining Intelligence, the latter two
enabling smarter contextual awareness and prediction of potential threats
leading to proactive prevention of them. The associated horizontal economic
implications of this evolution and its impact into the societal and economic
fabric are also tackled. Part of the results and analysis produced in this chapter
are the outcome of the work carried out in the FP7 EU project SafeCity, one of
the eight Use Cases of the FI Programme.
1 The Undergoing Safety Evolution in the Future Internet

A large urban growth has been recorded throughout the world within the last decade.
Such population increase in cities has implied an equally pressing demand for vital
public services such as transport, health, education or city security and citizens’
protection. Protecting citizens is one of the key factors for a government and also a
priority for the normal functioning of businesses, communities and civil society at
large. Safety in the cities is becoming more and more complex due to the constant
increase of city population and of city infrastructures complexity. In this sense, cities
and countries’ authorities have made a great effort in applying innovative approaches
and new technologies in the Public Safety domain in recent years, especially in order
The Safety Transformation in the Future Internet Domain 191
to reduce emergency response time and urban crime: for example, digital surveillance
cameras have been placed in many critical areas and buildings throughout cities and
call dispatchers have been created to distribute the emergency calls. Moreover,
advanced technological capabilities facilitate urban public safety systems to become
not just more interconnected and efficient, but also smarter and self-adaptive. Instead
of merely responding to crimes and emergencies after a critical situation, novel smart
systems emerge to analyse, anticipate and, actually, contribute to preventing them
before occurring. After the terrorist attacks of March 2004 in Madrid, the city
developed a new fully integrated Emergency Response Centre which, after an
incoming emergency call, simultaneously alerts the required emergency agency
(police, ambulance and/or fire brigade). The system can recognize if alerts relate to a
single or multiple incidents, and assign the right resources based on the requirements
coming from the ground. Furthermore, specialized video analytics systems are
successfully installed for traffic surveillance purposes. These are CCTV-based
systems capable of automatically detect illegal vehicles behaviour (e.g. cars stopped
in forbidden areas, going in the opposite direction), restricted entries behaviour (e.g.
bike entering in a forbidden road), stolen vehicles, etc. In addition, M2M
communications, that is, intelligent communications by enabled devices without
human intervention, are nowadays present in home and industrial security monitoring
systems and alarms. Several Public Safety organizations and Public Administrations
are using sensor networks to monitor environmental conditions or to be temporally
deployed driven by an emergency situation. Other advanced technologies are focused
on enhancing emergency notification mechanisms, fire and enforcement records
management, surveillance, etc.
As presented, outstanding capabilities offered by advanced technologies are
currently in use for safety purposes. However, there is still a wide list of non-satisfied
safety capabilities requested by Public Safety agencies. Several on-going initiatives
research upon how Future Internet can assist these entities in their daily work and
during emergency response phases. That is the case of SafeCity (Future Internet
Applied to Public Safety in Smart Cities) [1], an EU-funded project under the FP7 FI-
PPP programme which proposes to enhance the role of Future Internet by developing
smart Public Safety applications of high value. SafeCity aims at significantly
improving the implementation and up-taking of Future Internet services in this safety
field by 2015, leveraging the internet infrastructure as the bases of Public Safety
centred open innovation schemes. It is focused in situational awareness (i.e.
surveillance of public facilities, transport stations, energy facilities, roads, citizens in
the streets; environmental monitoring), decision-making tools in C2 centres, seamless
usage of ad-hoc communication networks temporarily deployed to support additional
demand communication capacity (e.g. due to a major plan event) and alerting
population mechanisms.
This paper presents the state-of-the-art and on going advances in these three vital
technological fields (Internet of things, Intelligent Video Analytics and Data Mining
intelligence) that are envisaged as fundamental pillars of the FI infrastructure in the
Public Safety domain. It further continues discussing and concluding on what the
economic implications of such technological advances for Safety purposes are.
192 R. Gimenez et al.
2 Internet of Things, the Billion of Billions Connected Devices

When Applied to Safety
The evolution from the City of today to the future “Smart and Safe” City will be
highly driven by the introduction of advanced digital and ICT technologies. The city
will be submerged by millions of simple and sometime tiny devices: sensors, meters,
actuators that will represent the city’s organs of sensing. All these millions devices
shall be connected through a capillary network reaching all the peripheral devices.
Like in the human body, there will be peripheral sensing organs connected through a
peripheral nervous system to transmit the collected data to the central nervous system
i.e. the brain. The brain is the set of command and control centers of the city; there, a
diverse set of applications resides, a part of which are dedicated to citizens’ protection
and city infrastructure safety. The theoretical model for IoT services in the Smart City
is the sensing – actuating infinite loop.
To realize and put in place such a complex command and control system, it is
necessary to be based on a standardized ICT reference architecture tackling data
networks connectivity and diverse IT application platforms interoperability. Such
reference architecture is shown in Figure 1.
Fig. 1. Architecture and infrastructure model for IoT services in a Smart City
The reference architecture consists of different layers: Devices Layer, Networks

Layer, Enabling Platform Layer and Applications Layer. The Devices Layer is
composed by many types of smart devices (meters, actuators, sensors, etc.) that have
both computation and data transmission capabilities. Many theoretical studies are
considering which can be the best frequency in the spectrum to be allocated for short
range RF devices. The GHZ frequency range seems not so suitable due to the very
short geographical coverage and to the high sensitivity to electromagnetic noise.
So the bands of frequency currently under consideration in Europe are: 169 MHZ and
868 MHZ. All the devices need to be managed by a M2M (Machine To Machine)
Platform with the following main features: open and standard interface with devices,
open and standard interface with applications, legacy, non standard, adapters, devices
and applications self discovery and identity management (access controls),
connectivity management (session, mobility), content management (QoS), security,
privacy and trust, service management (auto provisioning, auto configuration, self
healing, SW and FW upgrade, ...) for applications and devices, asset management
(SIMs Card for example), etc.
The Enabling platform layer besides M2M platform shall host also databases to
manage big quantity of data and data mining capability to extract “Meaning” from the
huge amount of data (see chapter 4). Capabilities of video analytics can also be part of
the enabling platforms as enabling capabilities for many application based on image
recognition (see chapter 3). Geographical localization of devices is also important to
intervene in the geographical area impacted.
Finally, the application layer is where the various applications reside. The
applications use web services APIs provided by the enabling platform layers. The
architecture is based on state of the art of web 2.0 techniques like [2]: Service-Oriented
Architecture (SOA), Software as a Service (SaaS), Mashups or Structured Information.
The SOA (OASIS Reference Model for SOA#) is an architectural paradigm for
integrated services available on the Net and owned and managed by different entities.
With SaaS, the software for implementing services is not locally installed and self
contained (an example of SaaS is world editor not installed on the computer where the
editing is done, but available in the Net). Mashup techniques are based on SOA and
enable to integrate different services to create a new and original service that can be as
well available in the Net for future mashups. Last but not least, XML family languages
have enabled the exchange of structured information between applications in the Net
without any previous design phase in the databases. Regarding data connectivity the
debate is open and research is on going to assess if public networks can be reusable for
safe city applications. The main consideration in favor of a re-use of current commercial
IP networks is that they are already in place while to build a specific network
infrastructure for the smart city case would require efforts, time and money that cannot
be spared (not to mention network planning and management).
Safety applications will leverage the IoT platforms described previously; in
particular the capillary network hot spots will be very important points for installation
on the territory of safety oriented sensors and actuators. First of all the IP Cameras
sending video streaming in Real Time can be managed as “Smart Things” both in
terms of data collection and in terms of operation management (maintenance in case
of faults). The capillary networks hot spots can also be the points of installation of
tools for alerting citizens. The alerting phase in safety services is very important.
When there is some emergency situation, citizens shall be informed as soon as
possible especially citizens close to the emergency areas. To alert citizens, Digital
Signage panels or totem can be installed in the capillary networks points. Moreover
through broadband connections it shall be possible to send alerting messages directly
to the mobile devices of the citizens in the Area or close to the area using for example
WiFi short range connections. To summarize the IoT is important for the safety smart
services and eventually safety smart services can defined as IoT services.
3 The Radical Change That Intelligent Video Analytics Is

Bringing to Safety in Smart Cities
A common problem in the operation of video surveillance systems is the sheer
volume of information: there may be thousands of cameras installed, and it is simply
impossible to physically monitor all of them on 24/7 basis. Furthermore, it is
relatively easy to “fool” an operator by innocuous-seeming appearance or behavior.
The wide deployment of digital video surveillance has led to computers replacing
human operators and security personnel in order to efficiently monitor and analyze
video footage in real-time and trigger alerts not affected by human fatigue and
distraction as in the case of human operators. Video Analytics (VA) is about the use
of dedicated software and hardware to analyze captured video and automatically
identify specific objects, events, behavior or attitudes in video footage in real-time.
Video analytics enables video surveillance to become a proactive monitoring tool that
signals the need for immediate intervention by guards, police, or other security
personnel. Video surveillance systems become thus more efficient being able to
automatically recognize situations and trigger alarms or other actions (such as door
locking). With video analytics acting as a remote observer (as shown in Figure 2),
security personnel may receive notifications of an intruder or other suspicious event,
and potentially act before a crime takes place – this added value brought around is
crucial for crime prevention and safer cities.
Fig. 2. Video Analytics in the landscape of a Smart and SafeCity
Given the explosion in the amount of video footage captured by security forces, the
need to develop automatic intelligent methods for detecting suspicious people, objects
or activities in video to trigger immediate alerts or further analysis has been widely
recognized. Intelligent video analytics tools have been emerging for that purpose,
deployed in the safety domain. However, recognizing objects and people in loaded
scenes, identifying a person based on gait, recognizing complex behaviors and
conducting analytics in multi-camera systems are still among the main challenges of
research in this field. In video analysis, a monitored scene is visually analyzed to
extract foreground, eliminating all irrelevant background information to the
understanding of the problem at hand. A large number of methods exist, including
adaptive Gaussian mixtures [3], which can be used together with shadow filters [4].
When medium or close field video streams are available (depth info), then more
sophisticated scene analysis can be provided, e.g. body shapes and regions can be
extracted. The dynamics of the evolving scene are interpreted and, according to the
density and clutter in the scene, it may be possible to track single persons or moving
objects, even in complex scenes. Multiple cameras with overlapping fields of view
allow for 3D tracking. Such methods are heavily based on the quality of features
detected (appearance, shapes etc.) and fail if image primitives are not reliably
detected. There are approaches that attempt to infer events without construction of
models. The detection of complex motion using motion models based on HMMs
(Hidden Markov Models) targets to detect abnormal events in complex scenes. Apart
from building models, the extracted information is used to recognize the event,
usually under assumptions/rules. Other methods achieve event recognition by relying
on both low-level motion detection and tracking, and high level recognition of
predefined (threat) scenarios corresponding to specific behaviors.
Advancements in video analytics technology have increased recognition abilities,
dramatically reducing false alerts due to weather, sun positions, and other
environmental factors. Some of today’s video analytics capabilities, along with safety
example cases that these may handle, include:
• Character (e.g., alphanumeric) and inscription recognition for reading license
plates, name tags, and containers; for e.g. suspicious parked cars detection.
• Facial recognition; for criminal/terrorist identification in public places
(metro, airport, large public squares, etc.)
• Density of people, people counts, behavior (such as loitering, fighting,
reading, sampling), slip-and-fall detection, gang activity, tailgating (vehicle
or human) in restricted areas, a person coming over a fence;
• Object removal and tracking; for e.g. theft detection cases
• Smoke detection; for potential fire detection
• Pattern recognition and directional motion;
• Tampering (such as with ATMs or other devices);
• Illegally parked cars, unattended bags, spills; for citizens’ protection
• Camera sabotage or malfunction, etc.; for crime intention detection
Intelligence and detection accuracy increases when one combines many of the above
capabilities together, or fuses detection results from the analysis of diverse
data/sensor inputs in an IoT infrastructure. For example, it is now possible to allow
entrance to a secure building by linking a fingerprint with a face and a person’s voice,
as well as personal handwriting, requiring all to match before granting access. Today's
intelligent video analytics systems can even spot potential problems by analyzing how
people move in multi-camera crowded scenes – many video streams are analyzed
simultaneously flagging suspicious people or objects, directing security personnel to
focus on particular activities. Artificial intelligence combined with video analytics
adds an intelligence layer, allowing learning of patterns while analyzing and dropping
false alarm rates. Finally, the use of both server-based (up to now the prevailing
architecture) and embedded, on-camera video analytics has led to even better
performance and lower energy and bandwidth consumption.
Nowadays, there is another great challenge to be faced due to the great demand for
respect for citizen’s privacy in order to retain public trust. The Anonymous Video
Analytics (AVA) technology has emerged for that purpose [5], which uses pattern
detection algorithms to scan real time video feeds, looking for patterns that match the
software’s understanding of faces. The data is logged and the video destroyed on the
fly – with nothing in the process recognizing the persons who passed in front of the
sensors. In safety applications, only the identity of suspicious people, logged in a
database, is found and revealed. The advantages of intelligent video analytics for
enabling safe cities as Future Internet applications in combination with other
technologies, such as sensor networks or data mining, fusion and decision support, are
thus numerous.
4 The Data Mining Intelligence in Smart Cities for Safety

Data mining has become the third key feature of many safety initiatives in smart
cities. Often used as a means for detecting fraud, assessing risk and product retailing,
public safety agencies can use predictive modelling and data mining techniques to
look for previously unknown valid patterns, relationships in large data sets and
process improvements in situational awareness and command centres. These smart
public safety systems can collect data from different processes, systems, and devices
and can apply intelligence to this mass of data. The intelligence applied to this data
can detect patterns of incidents and generate new insights, so that officials can make
well-informed decisions and take action in near real time. Also, using data mining
saves time for field personnel, reduces costs, and avoids the need for travel. Instead of
just reacting to crimes and emergencies, with these new technologies and capabilities,
public safety officials can perform analysis so that they can anticipate and work to
prevent incidents.
Specifically, public safety agencies like Richmond, Memphis and Edmonton Police
Departments apply data mining intelligence to tactical crime analysis in order to
review extremely large datasets and incorporate a vast array of variables, far beyond
what a single analyst, or even an analytical team or task force, can accurately review.
Also, data mining can be used to identify a crime or series of crimes associated with
an increased risk of escalation or violence; thereby, it is facilitated the apprehension
of the suspect and increased the possibilities of interrupting a crime series before
serious escalation occurs. Besides, data mining is considered as an essential technique
for analyzing Internet and Web log data since monitoring and characterizing “normal”
activity can help to rapidly identify unusual or suspicious activities in large datasets,
providing actionable patterns for use in subsequent analysis and surveillance. So,
public safety agencies can use it to identify and characterize extremely rare events,
anomalies, and patterns in relatively large datasets.
However, the continuous increase in data volumes causes great and challenging
difficulties in processing, analyzing and extracting valuable, new and useful
information for decision support tools. Therefore, methods for efficient computer-
based analysis are indispensable. In particular, support decision making can greatly
benefit from methodological techniques developed in the new interdisciplinary field
of Knowledge Discovery in Databases (KDD) [6] encompassing statistical, pattern
recognition, machine learning (ML), and visualization tools to support automatic data
analysis and discovery of regularities (patterns) that are implicitly encoded and
hidden within the data.
Besides mining knowledge from large amounts of data, annotation and correlation
of data from numerous and diverse digital evidence sources are essential in the
context of public safety. Annotation and correlation of data across multiple devices in
order to highlight an activity matching a scenario of interest are considered as a
promising technique to support the public safety agencies activities using a large
volume of information derived from heterogeneous environments. Therefore, there is
a need for normalization in the representation of data from multiple sources of digital
evidence in order to support such pattern recognition [8].
By providing a normalised view of all the data available, generating scenarios of
interest, mining of behavioural patterns and correlation between events can be
established. The needs for new architectures that incorporate techniques to analyse
data from multiple sets of digital evidence used by police and other investigation
entities and to represent such data in a normalized manner are presented in [7].
Currently, techniques based on semantics are applied for annotation and correlation
of data in the Safety and Security Knowledge Domain. Semantic data modelling
techniques provide the definition and format of manipulated data. They define
standardized general relation types, together with the kinds of things that may be
related by such a relation type. In addition, semantic data modelling techniques define
the meaning of data within the context of its interrelationships with other data. At this
point, it is where ontologies fit into, which are actually the semantic data models.
Ontology [9] is a formal representation of knowledge as a set of concepts within a
domain, and the relationships between those concepts. It is used to reason about the
entities within that domain, and may be used to describe the domain. Data models,
metadata and annotations, classification schemes and taxonomies, and ontology are
greatly used in a variety of applications and domains. In the security and safety
application (knowledge) domain, effective data modelling and knowledge
representation facilitate automated semantic-based analysis of large volumes of data
and identification of suspicious or alert situations and behaviours. Their added value
remains in sharing and extending such models and representations with other
stakeholders and similar applications to facilitate data interoperability and unified
reasoning within the same knowledge domain.
Finally, data management during disaster/crisis situations also requires facing the
same, mentioned above problems, due to the fact that disaster data are extremely
heterogeneous, both structurally and semantically. This creates a need for data
integration and ingestion in order to assist the emergency management officials in
rapid disaster recovery. Since the data in disaster management could come from
various sources and different users might be interested in different kinds of
knowledge, data mining typically involve a wide range of tasks and algorithms such
as pattern mining for discovering interesting associations and correlations; clustering
and trend analysis to classify events in order to prevent future reoccurrences of
undesirable phenomena. Due to the fact that real-world data in disaster management
tend to be incomplete, noisy, inconsistent, high dimensional and multi-sensory etc.,
development of missing / incomplete data correlation approaches in order to increase
the situational awareness can be especially beneficial in this context [10].
5 The Related Eco

onomic Impact in This Transformation
As is shown in Figure 3, one of the sectors where Future Internet stands out iin a
Smart City is Security. As A has been mentioned at the beginning of this repport;
‘protecting citizens is one of o the key factors for a government and also a priority for
the normal functioning of businesses,
b communities and civil society at large’. Iff for
instance, any facility in ch harge of providing day-to-day essential services sufferrs a
disaster (e.g. terrorist attacck); the service interruption should cause huge damagee to
society in the form of socio-economic losses, socio-political adverse effeects,
environmental consequencces or even substantial human casualties, each beeing
accompanied by related cossts. Anticipating and preventing those potential threats has
been widely analysed as an n essential aspect in order not only to keep the wellbeingg of
modern societies but also as a a cost-effective solution for any organization (publicc or
private) in charge of those infrastructures.
i
Fig. 3. The Smart City Framework
Moving security toward ds a most effective system will involve improvementss in

most of the other indicattors such as productivity, flexibility, labour markett or
budgetary savings. For exaample, introducing a new security architecture framew work
based on pre-processing seensors will enable clearer data arriving to first respondders,
shortening times in their actuations.
a This reduction shall provoke optimizationss in
times and resources, which h will reduce costs, rising higher productivity of secuurity
personnel, increasing job satisfaction and increasing efficiency and effectivenness
process. It is also worth mentioning
m that public savings and situation can also bbeen
extrapolated to business an nd companies features, provoking similar conditions, beeing
the impact in accordance with the business size and previous situation.
In the Internet of Thingss domain, ‘SmartCities’ will be submerged by millionss of
devices: sensors, meters, acctuators, connected through a capillary network reachingg all
the peripheral devices. Thiss process shall register several factors in real time, allow
wing
for instance a smart manag gement of potential threats or real emergencies allocatedd in
different places of the city. In relation to video analytics and data mining domain, iit is
also worth it to stand outt several examples of cities where cameras and senssors
deployed are already manaaged in a smart way (together with M2M technologies and
data mining) contributing to reduce possible emergencies as well as its response time.
For example, New York City for their City Control uses an innovative four
dimensional, integrated visualization technology that provides automated situational
awareness for anyone responsible for Securing and Protecting Infrastructure and/or
Human Assets. These technologies contribute on an essential way to optimize
capacity and first responders’ response time, both to beat to a potential risk, and to
response to an emergency. This phenomenon also brings an outstanding saving for
responsible organizations in charge of economic management of cities.
To conclude, it is worth to highlight that most of the benefits for the end users do
not create direct revenues, but significant operational savings and increased
efficiency. Also it is expected that transformation will produce significant economic
benefits for the society and business at large.
6 Conclusions
Smart Cities of tomorrow will provide a larger number of innovative services and new
capabilities that will highly contribute to reinforce the feeling of safeness in citizens.
Enhanced M2M communications will allow the massive usage of heterogeneous
sensors (smart meters) around the city and its surroundings, internet-connected and
self-configured devices that enable web-sensors access and surveillance-information
sharing among diverse safety agencies involved. Robust intelligent video analytics
that enable smarter contextual awareness will be applied not only for traffic purposes
but also for other aspects as suspicious objects/behaviors early detection, and will
represent the required answer to the existing explosion of video footage captured by
security forces who want to enlarge the automated detection capabilities of their video
surveillance systems. Predictive modeling and data mining techniques applied to
surveillance data enable the early detection of incidents and the generation of new
insights that efficiently support decision-makers. Depicted expected technological
advances within these three pillar areas clearly benefit Public Safety services with
intelligent real-time surveillance capabilities, efficient early detection mechanisms,
enhanced information visualization and sharing, and semi-automatic decision support
systems at Command and Control centers. Public Authorities will extremely reduce
the response time to emergencies (see that Madrid Emergency Response Centre
helped to reduce it to 25%) since innovative internet-based capabilities are expected
soon, for instance, an efficient monitorization for road safety purposes detecting
drastic weather changes, road condition, foreign objects, or the early detection
mechanisms based on video analytics of suspicious/missing people, suspicious
behaviors, illegal entries, suspicious objects, etc., which can be even more efficient
with alerting capabilities to specific geo-graphically based population.
All these new techniques will have an important impact and fostering of economic
sustainability within a Smart City while offering high quality Public Safety services.
References
1. SafeCity Project: FI-PPP (Future Internet Public Partners Partnership) Programme (2011-
2013), http://www.safecity-project.eu/
2. Nuckull, D., Hinchcliffe, D., Governor, J.: Web 2.0 Architectures. O’Reilly (2009)
3. Stauffer, C., Grimson, W.E. L.: Learning patterns of activity using real-time tracking.
IEEE Trans. Pattern Analysis and Machine Intelligence 22(8), 747–757 (2000)
4. Cucchiara, R., Grana, C., Neri, G., Piccardi, M., Prati, A.: Sakbot system for moving
object detection and tracking. In: Video-Based Surveillance Systems-Computer Vision and
Distributed Processing, pp. 145–157 (2001)
5. Cavoukian, A.: Anonymous Video Analytics (AVA) technology and privacy. White Paper,
Information and Privacy Commisioner, Ontario, Canada (April 2011)
6. Han, J., Kamber, M.: Data Mining: Concepts and Techniques. The Morgan Kaufmann
Series in Data Management Systems, Jim Gray Series Editor. Morgan Kaufmann
Publishers (2000)
7. Osborne, G., Turnbull, B.: Enhancing Computer Forensics Investigation through
Visualisation and Data Exploitation. In: International Conference on Availability,
Reliability and Security, ARES 2009 (2009)
8. Mohay, G.: Technical Challenges and Directions for Digital Forensics. In: SADFE 2005:
Proceedings of the First International Workshop on Systematic Approaches to Digital
Forensic Engineering, p. 155. IEEE Computer Society, Washington, DC (2005)
9. Guarino, N.: Formal Ontology and Information Systems. In: 1st International Conference
on Formal Ontology in Information Systems (FOIS 1998), Trento, pp. 3–15 ( June 1998)
10. Hristidis, V., Chen, S.C., Li, T., Luis, S., Deng, Y.: Survey of Data Management and
Analysis in Disaster Situations. Journal of Systems and Software 83(10) (October 2010)
FSToolkit: Adopting Software Engineering Practices
for Enabling Definitions of Federated Resource
Infrastructures
Christos Tranoris and Spyros Denazis
University of Patras, Greece

[email protected], [email protected]
Abstract. Today organizations own resources and infrastructures

(i.e. networking devices, gateways, wireless devices) that would like to either
offer through the cloud model or to combine with resources of other
infrastructures. Federation can be enabled by means of a resource broker that
matches customer’s requested services and providers’ resources according to
the agreed SLA. Users need ways to define complex deployments and request
for resources without knowing the underlying infrastructure details. In this
paper we present the Federation Scenario Toolkit (FSToolkit) that enables the
definition of resource request scenarios, agnostic in term of providers. This
work adopts Software Engineering practices considering the concepts of
modeling and meta-modeling to define a resource broker and to specify
scenarios by applying the Domain Specific Modeling (DSM) paradigm.
FSToolkit is developed for experimentally driven research for validating
through testing-scenarios new architectures and systems at scale and under
realistic environments by enabling federation of resources.
Keywords: Federation, experimentally driven research, Resource Broker,

Domain Specific Modeling.
1 Introduction
Future Internet research needs new infrastructures for supporting approaches that
exploit, extend or redesign current Internet architecture and protocols. During the last
few years experimentally driven research is proposed as an emerging paradigm for the
Future Internet on validating through testing-scenarios new architectures and systems
at scale and under realistic environments. Until recently, testbeds used in testing
activities have usually a certain scope of testing capabilities. Organizations own
resources and infrastructure (i.e. networking devices, gateways, wireless devices) that
would like to either offer through the cloud model or to combine with resources of
other infrastructures in order to enable richer and broader experimentation scenarios.
Experimentally driven research addresses the need to evolve the test beds into
coherent experimentation facilities. This is possible by enabling large-scale federated
infrastructures of exposed organizational resources and testbed facilities. Such future
experimental facilities are leaded by global efforts like GENI [1] and FIRE [2].
202 C. Tranoris and S. Denazis
Federated infrastructures in experimentally driven research need models,

architectures and tools to address the definition and execution/operation/control of the
experiment. In our previous work [3], we presented a paradigm called Federation
Computing where it deals with the aspects of defining and operating/controlling
experiment scenarios or so called Federation Scenarios. We applied these concepts in
the context of the Panlab project [4]. A Federation Scenario is a well-defined
specification of (heterogeneous) services or resources and their configurations,
offered by a diverse pool of organizations in order to form richer infrastructures for
experimentally driven research. A Federation Scenario is the equivalent of an SLA
required by the end-user, which is the customer of the federation. These federation
scenarios represent customer needs such as i) evaluation and testing specifications of
new technologies, products, services, ii) execution of network and application layer
experiments, or even iii) complete commercial applications that are executed by the
federation’s infrastructure in a cost-effective way.
Fig. 1. Federations of Resource Providers and Brokers
A Federation Scenario describes end-user needs for services provided by resources

of a federated infrastructure. At this point resource brokers play a key-role in creating
and supporting federated infrastructures. A resource broker matches customer
requested services and provider resources from the federation. Federation and
resource brokers are well addressed by the cloud computing community in [5]:
“Federation is the act of combining data or identities across multiple systems.
Federation can be done by a cloud provider or by a cloud broker. A broker has no
cloud resources of its own, but matches consumers and providers based on the SLA
required by the consumer. The consumer has no knowledge that the broker does not
control the resources.”
Figure 1 displays resource providers forming a federation where a resource broker
is capable of exposing resources R to end-users in a uniform manner to create richer
infrastructures. Resource providers must have an API that exposes resources and
enables brokers to browse and manage them. The end-user can create scenarios
involving resources by directly going to a resource provider or by going to a resource
broker of a federation.
This work discusses how we adopted Software Engineering practices of Domain
Specific Modeling (DSM), where the systematic use of textual or graphical Domain
Specific Language (DSL) is involved. A DSL is defined as a specification language
FSToolkit: Adopting Software Engineering Practices for Enabling Definitions 203
that offers, through appropriate notations and abstractions, expressive power focused
on, and usually restricted to, a particular problem domain. For the language definition
an abstract syntax (the meta-model), a concrete syntax and semantics are needed. All
of these are captured in a solution workbench, which in this case is Eclipse, used both
as a development but also as a deployment environment.
Having stated the above, we present a meta-model for defining a resource broker
and how Domain Specific Modeling (DSM) is used to define Federation Scenarios.
We implemented a meta-model that describes resource brokers offering (representing)
services later mapped to resource providers. The Domain Specific Languages (DSLs)
used by resource brokers, resource providers and experimenters, have the proposed
meta-model as an abstract syntax. The meta-model is called Office meta-model, since
it has inherited its name from the Panlab Office which is used to federate resources.
However, the Office meta-model is generic enough to describe any resource broker.
A DSL called Office Description Language (OfficeDL) is used by resource brokers
or resource providers to describe them. The end-user (an experimenter or customer)
uses the Federation Scenario Description Language (FSDL). FSDL is a DSL to
describe the needed services of an experiment over a federated infrastructure. We also
discuss how we used Model-to-Model transformation between resource brokers in
order to import in the language heterogeneous resources by other resource brokers or
resource providers expressed with other models. Model-to-Text transformations are
used to generate wrapper code for exposing resources and for targeting different
provisioning engines.
The paper is structured as follows: First we present the proposed meta-model and
its core entities. Then we present the OfficeDL used by resource brokers and resource
providers and then we provide details of the FSDL and its concrete syntax in
describing Federation Scenarios. All the languages are supported by the FSToolkit
tooling which is also presented.
2 The Meta-Model Describing Resource Brokers and

Federation Scenarios
A Federation Scenario describes customer needs for services over a federated
infrastructure. To support this, we needed first to define a resource broker. Thus, we
define a meta-model the Office meta-model (figure 2, level M2) which describes
resource broker models (in M1) and eventually instantiations of them in Federation
Scenario definitions (M0). In the Office meta-model the core entity Office is defined.
An Office is a resource broker offering services and matching services and resources,
maintains users, and in general support federation scenarios.
In our work we define Offered Services and Offered Resources in Office as
follows:
An Offered Service is an abstract entity and it describes an offering along with its
configuration attributes, e.g. Computing Resource with memory, disk space, etc.
An Offered Resource is an entity that implements an Offered Service. e.g. Resource
Acme.Comp1234 is a resource of the provider Acme capable of implementing a
service of Computing Resource (creating Virtual Machines).
Fig. 2. The Office meta-model levels
An Offered Resource is supposed to be managed by Create-Read-Update-Delete

operations. So an Offered Resource is currently a really simple entity with a few
attributes exposed to the end-user. The same applies to an Offered Service.
The Office meta-model, is defined in Ecore: a variant of OMG’s MOF [6] that has
been developed in the Eclipse Modeling Framework [7] and is more or less aligned on
OMG's Essential MOF (EMOF).
The Office meta-model defines related entities and their relationships, what an
Offered Service is, what an Offered Resource is, how an Offered Service is supported
by a resource of the federation, taxonomies, service compositions, SLAs, users, etc.
Part of the meta-model is illustrated in figure 3, where it displays that an Office is an
aggregation of Offered Services, Users and Requested Federation Scenarios. The
Office aggregates Requested Federation Scenarios where an SLA (not shown) is
created for each one of them. Since the entity Office describes actually a resource
broker, it has an aggregation of providers offering resources. A Resource Provider is
viewed as a user of the Office. A Resource Provider has an aggregation of Sites and
eventually a Site contains the Offered Resources.
An Office matches Offered Services and Offered Resources. Having this, the Office
maintains some contracts the ResourceServiceContracts (see Figure 3 right side). A
contract helps the broker to match a service to a resource. From figure 3, one can see
that a ResourceServiceContract is between an Offered Service and an Offered
Resource. Some extra characteristics of the contract are described in the Availability
of the Resource and potential Cost.
Figure 4 displays part of the Office meta-model which is used as the abstract
syntax of the FSDL language. Classes here are instantiated later on while the end-user
specifies the Federation Scenario. The RequestedFederationScenario contains user
Credentials, ScheduledPlans, Import for URIs and most importantly a
ServicesRequest. The ServicesRequest is a composition of ServiceRequest, the
services that the end-user wants for his scenario. Each ServicesRequest references an
(Offered) Service and contains some requested ServiseSettingInstances.
Fig. 3. Part of the Office meta-model and ResourceServiceContract
Fig. 4. A view of the RequestedFederationScenario
3 The OfficeDL: Describing Resource Brokers and Resource

Providers
To enable rapid prototyping of the Office meta-model and to enable easy
instantiations of the meta-model we developed a DSL called OfficeDL. OfficeDL has
as abstract syntax the meta-model. The concrete syntax is based on the textual
modeling framework (TMF) of Eclipse [8] and specifically the Xtext framework [9.
An example of describing an Office called myBroker is given below:
Office "myBroker" {
registeredUsers {
OfficeCustomer "Tranoris" {
address "Address"
hasAccount Account "Name" {
password "Password" username "Username"
}
},
ResourcesProvider "ProviderA"{
offeredSiteList {
Site "WCL" {
ptm PTM "uopPTM" { IP "150.140.184.234" }
igwlist { IGW "uopIIW" { IP "150.140.184.231" } }
locatedAt SiteLocation "loc" {

address "Rion, Patras"
geocoords "5435345.00, 325345.00"
}
offeredResourcesList {
OfferedResource "UpatMI" {},
OfferedResource "VM_STANDARD" { },
OfferedResource "VM_MEDIUM" {},
OfferedResource "VM_LARGE" {}
}
}
}
},
The language tokens are with bold fonts and variables with other fonts. Having this, while
someone uses the language, he creates a model of his own office, defining: users, offered
services, resource providers, offered resources, etc… Some benefits of creating such a
DSL: there is a way to quickly check the meta-model for its correctness; tools can import
the instantiated models which are validated from the framework; resource brokers can use
it to describe their users, offered services, providers and contracts; finally, resource
providers may use it for describing only their own organization resources for local usage
and offer all the available tooling to their users.
It is expected that OfficeDL will be used for small to medium broker and provider
descriptions. For large organizations a permanent repository supporting the model is
more adequate. These descriptions though will be useful later on, when end-users use
them for defining their federation scenarios.
4 FSDL: A DSL for the End-User

The previous section discussed the OfficeDL which is used by a resource broker and
resource providers for describing federation entities. We have created another DSL
for enabling the end-user describing federation scenarios. The language is called
Federation Scenario Description Language (FSDL). In the simplest usage an FSDL
definition starts with the keyword RequestedFederationScenario followed by a name.
A set of import office statements that contain definitions of the offices (the resource
brokers, services and resources) may follow. Next, one can define either a resource
agnostic scenario request or specific resources of providers. To illustrate the above we
will discuss some examples.
The following, discusses a resource agnostic scenario request example (with a
request for offered services). The request is towards a broker brokerOfficeXYZ. We
would like to use an echo service that the brokerOfficeXYZ provides. The request is
described in the following FSDL :
RequestedFederationScenario myScenarioName
import office "http://brokerOfficeXYZ.org/myresourcedef.office";
RequestServices{
Service "brokerOfficeXYZ.echo" as myecho settings {//An echo
resource. Write something in input. Read the output to get it
Setting "input" : myinput = "Hello" //An input text for echo
Setting "sleeptime_ms" : mysleeptime_ms = "3000" //delay of echo
in msecs
}
}
Inside the RequestServices section we describe the request for services and their
initial settings. The keyword Service declares a new service request followed by the
name of the requested service. In the presented example we request the echo service
echo. After the as keyword we define an alias of the service (i.e. myecho). After the
settings keyword follows the section with the initial settings of the requested service.
In our example we define the two settings input (the input setting will be the output of
the echo service) and sleeptime_ms (delay of the message).
In the next example we present the case of selecting resources from specific
providers, where we use a slightly different syntax of the language. In this case, we
would like to use an echo resource that provider ProviderAcme offers. We have two
ways to express this request in FSDL. The first:
RequestServices{
Service "brokerOfficeXYZ.echo" as myecho offered by
"brokerOfficeXYZ.ProviderAcme" settings{
Setting "input" : input = "Hello" //An input text for echo
Setting "sleeptime_ms" : sleeptime_ms = "3000" //delay of echo
in msecs
}
}
The keyword offered by is used to define that the end-user wants to request the
resource by the ProviderAcme provide. Another way for expressing this request is as
follows:
RequestInfrastructure {
Resource "brokerOfficeXYZ.ProviderAcme.site.echo_rp12_or10782" as
myecho settings {
Setting "output" : output = "" //
Setting "input" : input = "Hello" //
Setting "sleeptime_ms" : sleeptime_ms = "2000" //
}
}
The RequestInfrastructure is used to describe a concrete infrastructure of resources
and their attributes by specific resource providers. Both approaches could be used for
different needs. Usually service definitions are more generic and contain generic
settings that all resource providers supply. However it is possible that a resource can
have more settings than the offered service it matches. The latter description of
describing the infrastructure is submitted for provisioning. In general, the section
ServicesRequest contains a list of ServiceRequest entities. The user creates instances
of ServiceRequests in the language referenced by the imported model. The syntax for
requesting an Offered Service is as follows:
Service "NAME_OF_SERVICE" as nameAlias([1.. numOfServices
])?(offered by "ResourcesProvider" (optional)? )? settings {
Setting "NAME_OF_SETTING":settingNameAlias (= staticValue)?
(assign +=SettingInstance|STRING] ( , SettingInstance )?
Setting "NAME_OF_SETTING":settingNameAlias (= staticValue)?
(assign +=SettingInstance|STRING] ( , SettingInstance )?
...
...
}
Where:
• NAME_OF_SERVICE: a full qualified name of the service
• nameAlias: a user chosen value to name the service followed optionally by how
many services he wants
• offered by is optionally to indicate to the broker that we need the specific provider.
• the optional keyword says to the broker to try to match the selected provider if
possible
• NAME_OF_SETTING: the name of an attribute of an offered service
• settingNameAlias: a user chosen value to name the setting. If after the alias there is
a = then the setting can have a static value. If there is the keyword assign the user
can assign the value of another setting.
A more complex example to illustrate FSDL is the following: An end-user wants to
deploy a XEN VM image to 15 machines. The resource broker brokerOfficeXYZ will
allocate these later to his resource providers. The FSDL specification is as follows:
RequestedFederationScenario deployingAXenImage
RequestServices{
Service "brokerOfficeXYZ.xenimagestore" as myXENImageP2ner settings{
Setting "Name":imgname = "myXENImageP2ner"
Setting "InputURL":inputurl
="http://196.140.184.233/myxenimage.img"//The url to copy from
Setting "OutputURL":outputurl //holds the location of the stored
image, to be used by testbed's resources
}
Service "brokerOfficeXYZ.xenvmdeploy" as clients[1..15] settings{

Setting "CAP": cap = "50"
Setting "MEM": mem = "512"
Setting "URL": url assign "myXENImageP2ner.outputurl"
Setting "NAME": name = "client"
}
}
The user wants 2 services. The xenimagestore is used to move a XEN VM image to
be used by a XEN host, where the InputURL setting defines the source of the image.
The xenvmdeploy service is responsible for deploying the XEN image to a computing
resource. Some parameters are depicted in the example. Also the keyword assign is
used when we want to assign as input to this setting the value of another setting by
another offered service. The clients are declared as a group of services
(clients[1..15]). This gives the end-user flexibility when later runs the scenario
to execute commands on all the services (and eventually the resources) of the group.
Each ServiceRequest contains a list of settings that the user can define for the
scenario. The end-user can either define values for each setting (eg. an integer or a
string) or can assign output values from other resources of the scenario.
To help the end-user with the syntax and protect from syntax errors the FSToolkit
[15] environment has a specific FSDL editor. The editor is based again on the textual
modeling framework (TMF) of Eclipse and is installed by the end-user as Eclipse
plugins. Figure 5 displays an overview of the FSToolkit with installed FSDL plugins.
It contains views that help the end-user during the description of a scenario. On the
left hand side there are views to see user projects, Offered Services from available
Offices and stored scenarios on those offices. On the middle we depict the editor of
FSDL files. The editor is capable of making syntax validation and the context-assist
Fig. 5. An overview of the FSToolkit Federation Scenario editor
utility helps the end-user with the correct syntax by suggesting commands, keywords
and variables. Moreover, double clicking on offered services triggers automatic text
injection in the scenario description.
5 Towards Provisioning of Resources

Provisioning resources in a federated infrastructure is also a necessary step during a
scenario’s lifecycle. As discussed, our model assumes that resource brokers offer
services that later on are matched to resources via contracts under certain availability,
cost, policy, etc. Using FSDL, the end-user requests offered services from resource
brokers.
All this “contract-oriented” information is used by a module called Resource
Advisor, which transforms the Federation Scenario into a detailed list of requirements
for specific resources. The Resource Advisor proposes to the Federation Scenario
developer different Implementation Plans to continue, under certain cost and
availability of the resources. In this way we have created a model of an SLA for
federation scenarios in order to assign responsibilities to a certain resource for every
item contained in an SLA. An SLA aggregates contracts for each requested service.
To this end, a provider’s resource is responsible for a specific requirement of the
SLA. This approach of contracts and responsibilities of resources helps also towards
monitoring an SLA for different aspects (ie metering, service quality, security, etc).
The Resource Advisor module is a plugin in the FSToolkit environment.
The provisioning workflows invoke RESTful commands towards Broker Gateways
(BGW) and eventually provision provider resources. A similar process is followed for
tear down and releasing the resources.
6 Provisioning/Controlling Resources of Federation Scenarios:

The Federation Computing Interface
What is critical with the operational part of a federation scenario is the proper and
valid configuration of the participating resources. While the scenario is operated by a
customer (i.e. during an application deployment or during an experiment on the
federated infrastructure), the federation must ensure that all SLA terms are fulfilled
and nothing is violated or falls out of the scope of the SLA. To this end, the SLA must
be constantly monitored for different aspects (i.e. metering, service quality, security,
etc). In [3] we presented some initial aspects of such an API, which is called
Federation Computing Interface (FCI) [14].
Fig. 6. A plugin based architecture of FSToolkit and Extension Points
7 Extending the FSToolkit via Extension Points

FSToolkit is based on the Eclipse platform and is being deployed to end-users as a set
of plugins. Additionally, FSToolkit can be extended through defined Extension
Points. Figure 6 shows this concept. There are three main Extension Points in
FSToolkit. The IWoklflowEngine and the IProvisioningJobListener are used by
plugins that are capable of handling provisioning of resources. The third extension
point, the IOfficeRepository, can be used by resource providers and brokers to expose
resources to the end-users, in order to create federation scenarios. A provider in order
to support provisioning of resources, the extension point IProvisionResource of the
Provisioning Engines must be implemented.
8 Conclusions and Future Work

This paper discusses how we applied Software Engineering practices and especially the
Domain Specific Modeling paradigm, for defining federation scenarios. A meta-model
for resource brokers and resource providers was presented. Moreover we developed a
family of DSLs targeting brokers, providers and end-users having the meta-model as
abstract syntax. All appropriate tooling supporting is given through FSToolkit. All
presented tools are licensed under the Apache License, Version 2.0. The meta-model
can be downloaded from http://svn.panlab.net/PII/repos/Software/sources/FCI/org.
panlab.software.office.model/model/. More details, instructions, source code and
downloads are available also at our web site http://nam.ece.upatras.gr/fstoolkit.
Acknowledgments. The research leading to these results has received funding from
the European Union's Seventh Framework Programme (FP7/2007-2013) from project
PII- Panlab and under grant agreement n° 287581 – OpenLab.
References
1. National Science Foundation, GENI, http://www.geni.net (last accessed February
12, 2012)
2. European Commission, FIRE website,
http://cordis.europa.eu/fp7/ict/fire (last accessed February 12, 2012)
3. Tranoris, C., Denazis, S.: Federation Computing: A pragmatic approach for the Future
Internet. In: 6th IEEE International Conference on Network and Service Management
(CNSM 2010), Niagara Falls, Canada (2010)
4. Website of Panlab and PII European projects, supported by the European Commission in
its both framework programmes FP6 (2001-2006) and FP7 (2007-2013),
http://www.panlab.net
5. Opencloudmanifesto, Cloud Computing Use Cases White Paper,
http://www.opencloudmanifesto.org/ (last accessed February 12, 2012)
6. OMG website. Catalog of OMG Modeling and Metadata Specifications,
http://www.omg.org/technology/documents/modeling_spec_
catalog.htm (last accessed February 12, 2012)
7. Steinberg, D., Budinsky, F., Paternostro, M., Merks, E.: EMF eclipse modeling
framework, 2nd edn. Addison Wesley (2008)
8. Eclipse Foundation website, http://www.eclipse.org (last accessed March 27,
2011)
9. Xtext framework website, http://www.eclipse.org/Xtext/ (last accessed
February 12, 2012)
10. RADL, Panlab wiki website, http://trac.panlab.net/trac/wiki/RADL (last
accessed February 12, 2012)
11. Teagle, http://www.fire-teagle.org (last accessed February 12, 2012)
12. Specification Business Process Execution Language for Web Services, Version 1.1,
ftp://www6.software.ibm.com/software/developer/library/
ws-bpel.pdf (last accessed February 12, 2012)
13. Belaunde, M., Falcarin, P.: Realizing an MDA and SOA Marriage for the Development of
Mobile Services. In: ECMFA: European Conference on Modelling Foundations and
Applications, pp. 393–405 (2008)
14. Federation Computing Interface (FCI), Panlab wiki website,
http://trac.panlab.net/trac/wiki/FCI (last accessed February 12, 2012)
15. Federation Scenario Toolkit (FSToolkit) web site,
http://nam.ece.upatras.gr/fstoolkit (last accessed February 12, 2012)
NOVI Tools and Algorithms for Federating Virtualized
Infrastructures
Leonidas Lymberopoulos1, Mary Grammatikou1, Martin Potts2, Paola Grosso3,

Attila Fekete4, Bartosz Belter5, Mauro Campanella6, and Vasilis Maglaris1
1
National Technical University of Athens
2
Martel Consulting
3
Universiteit van Amsterdam
4
Eötvös Loránd Tudományegyetem
5
Poznan Supercomputing and Networking Center
6
Consortium GARR
Abstract. The EC FP7/FIRE STREP project NOVI - Network Innovation over

Virtualized Infrastructures - explores efficient approaches to compose
virtualized e-Infrastructures towards a holistic Future Internet (FI) cloud
service. Resources belonging to various levels, i.e. networking, storage and
processing are in principle managed by separate yet inter-working providers. In
this ecosystem NOVI aspires to develop and validate methods, information
systems and algorithms that will provide users with isolated slices, baskets of
resources and services drawn from federated infrastructures. Experimental
research accomplished thus far concludes the first phase of NOVI, with early
prototypes of semantic-aware advanced control & management plane
components being deployed and tested. The NOVI testing environment is based
on combining PlanetLab and FEDERICA, two dissimilar virtualized
experimental infrastructures with attributes widely anticipated in a FI cloud.
This federated testbed is stitched at the data plane via the NSwitch, a distributed
virtual switch developed within NOVI.
Keywords: NOVI, Future Internet, FIRE, Virtualization, Federation.
1 The NOVI Project: Goals and Objectives

We report in this paper experimental work within the NOVI FIRE project [1] towards
a semantic-aware control and management plane for federating heterogeneous
virtualized infrastructures and for establishing data plane connectivity amongst virtual
resources offered by separate virtualized infrastructures. The goal is to offer
automated advanced capabilities to users of the federation: intelligent resource
mapping, policy-driven access and resource allocation, context aware resource
discovery, transparent data plane connectivity and monitoring of combined user slices
and substrate resources across domains. Experimental tool development and
validation are performed in a testbed environment, based on two dissimilar virtualized
infrastructures: FEDERICA [2] and PlanetLab [3]. The former, partially supported by
the EC FP7/Capacities Programme, provides users with a combination of Virtual
214 L. Lymberopoulos et al.
Machines and Logical Routers, interconnected via Layer 2 VLAN technology

extended over dedicated circuits provisioned by European National Research &
Education Networks (NRENs) and GÉANT [4]; the latter is a popular experimental
infrastructure, partially supported by the US NSF GENI Programme [5], that offers
collections (slices) of virtual computing resources (slivers) within more than a
thousand hosts, globally distributed over the legacy Internet. The selection of these
platforms provides NOVI with a combined testbed, exposing experiments to a wide
range of attributes as expected in a Future Internet federated cloud: FEDERICA
combines commercial virtualization tools providing virtual machines and logical
routers, interconnected with gigabit controlled connectivity; PlanetLab offers a highly
distributed virtual machine selection interconnected over the existing public Internet,
thus presenting distributed applications experiments with actual connectivity
limitations and unpredictable wide area networking behavior. In conclusion, the
combined PlanetLab and FEDERICA testbed for NOVI’s experimental research
captures basic features applicable in federated heterogeneous environments that are
expected to serve a wide range of user communities.
The paper is organized as follows. Section 2 presents our work on a domain-
independent Information Model aiming to capture the main abstractions of shared
resources and services within a NOVI federation. Section 3 presents the main
components of NOVI’s federated control and management plane and provides an
overview of NOVI’s distributed virtual switch (NSwitch) for data plane stitching.
Section 4 presents the combined testbed for NOVI’s prototype deployment and
experimentation. Finally, section 5 concludes the paper and provides directions for
future work in the remainder of the project.
2 NOVI Information Model

An agreed-upon Information Model (IM) provides consistent and shared semantics
and descriptions of available resources and services in a federated environment. In
NOVI we developed a novel IM and the associated data models as existing IM efforts,
listed in NOVI Public Deliverable D2.1: Information Models for Virtualized
Architectures [6] did not cover our two-fold objective: (a) to support the modeling
abstractions to cater for a federation of infrastructures, e.g. the FEDERICA and
PlanetLab platforms of the NOVI’s testbed; (b) to include the necessary concepts so
that can be used to model other Future Internet (FI) infrastructures that could
participate in a NOVI-like federation.
We fully embraced a Semantic Web approach and defined data models using the
Web Ontology Language - OWL [7]. This choice has been driven by the desire to
support reasoning and context awareness, which in turn allow NOVI to create
efficient and complex services with resources available within the federation.
The NOVI IM consists of three distinct but related ontologies; this modular
approach is chosen on purpose to make the model more easily usable outside the
project by parties interested in specific aspects. The NOVI IM defines a resource
ontology, a monitoring ontology and a policy ontology.
The Resource Ontology provides the concepts and methods to describe the
resources offered by Future Internet platforms and how they are connected together in
a federated environment. This ontology provides the basis for topology and request
NOVI Tools and Algorithms for Federating Virtualized Infrastructures 215
descriptions and the terminology for describing physical nodes, virtual nodes, virtual
topologies, etc. The Resource Ontology supports the operation of all the services of
NOVI’s Federated Control & Management Architecture, which will be presented in
the Section 3 of this paper. For example, it is used to express requests within the
NOVI GUI or by the Resource Information Service and the Intelligent Resource
Mapping Service to communicate when coordinating the exchange of information
about resources suitable for the embedding of virtual resources. The Monitoring
Ontology extends the Resource Ontology to provide descriptions of the concepts and
methods of monitoring operations, such as details about monitoring tools, how these
relate to the resources, types of measurements that can be gathered etc. This ontology
provides the primary support to the operation of the Monitoring Service. Finally, the
Policy Ontology also extends the Resource Ontology by providing descriptions of the
concepts and methods for the management and execution of policies defined within
member platforms of a NOVI federation. This ontology supports the operation of the
Policy Service. More information on the developed ontologies can be found in the
project’s public deliverable D2.2: First Information & Data Models [8].
3 NOVI Federated Data, Control and Management Plane

Architecture
NOVI’s novel algorithms, methods and services are initially based on the Slice
Federation Architecture - SFA [9] as developed for the PlanetLab control &
management plane federation. In SFA, a resource specification - RSpec is an XML-
file describing resources bound and available to a user slice in terms of hardware
characteristics, network facilities, constraints and dependencies on their allocation.
NOVI extends SFA with advanced context-aware federation mechanisms (intelligent
resource allocation, monitoring, policy management and virtualized resources
discovery) and automating slice control & management operations anticipated within
a complex NOVI federation.
The high level overview of the NOVI Data, Control & Management (C&M)
architecture is shown in Fig. 1. It consists of three different layers:
1. At the bottom layer heterogeneous platforms (domains, infrastructures) contain
the virtual resources to be allocated to user requests for combined slices. Data
plane connectivity within a NOVI federated slice is achieved using NOVI’s
Distributed Virtual Switch – NSwitch
2. The middle layer components are used to provide basic C&M federation
capabilities across platforms. In the figure we depict implementation choices
referring to SFA (e.g. cross-domain authentication via synchronized registries
and user-specified slice operations)
3. The top layer implements NOVI C&M services that aim at offering advanced
capabilities to the federation users (e.g. intelligent resource mapping, policy-
driven access and resource allocation, context aware resource discovery,
transparent monitoring of combined user slices and substrate resources across
domains). It leverages federation mechanisms of the middle layer (SFA),
complementing them with advanced C&M functionality.
Fig. 1. NOVI Federated Data, Control & Management Architecture
For each infrastructure (platform) in the federation, as demonstrated within the

NOVI testbed, separate NSwitch, SFA and NOVI C&M instances need to be
deployed. In what follows we outline functionality of components within the latter.
3.1 NOVI API

The NOVI API provides the entry point for interacting with NOVI C&M services. It
has three main tasks: (1) Accept requests from authenticated users containing
resources requirements represented in NOVI Information model; (2) Handle and
deliver the request to the appropriate component within NOVI Service Layer; (3)
Provide user feedback on how their request is handled before the experiment starts
being executed in its related NOVI slice.
NOVI Tools an
nd Algorithms for Federating Virtualized Infrastructures 217
Fig. 2. Role of the NOVI API
As shown in Fig.2, the NO OVI API receives requests from the NOVI GUI. The GU UI is
based on the Ontology Instan nce Editor - OIntEd [10], which was originally used to asssist
in the development phase of o the NOVI IM and subsequently was customized to alllow
users to create and send requ
uests for NOVI slices. In its current implementation, the NOOVI
GUI provides an intuitive draag-and-drop interface for this instantiation process and alloows
users to define relations bettween instantiated objects. For example, a user can definne a
virtual network topology aloong with the characteristics for requested resources. For evvery
request, the GUI generates an n OWL document based on the NOVI IM which is sent too the
NOVI API by means of an HTTPH post request.
Using the NOVI GUI (accessible
( online at http://novi-im.appspot.com/) the uuser
can choose from the availaable ontologies in order to define the topology of the slice
for his experiment.
3.2 Resource Informatiion Service(RIS)

The Resource Information Service (RIS) acts as a single point of contact for otther
NOVI services to acquiree information about the status of virtual and substrrate
resources. Resource disco overy encompasses locating and retrieving informattion
across the federated virtuaalized substrate network in a decentralized way witth a
R uses the Request Handler to communicate with the
scalable query process. RIS
underlying platform, to reseerve resources and to obtain the resource advertisementss. It
uses the Monitoring Servicee to query on the availability and the status of the resourrces
and the Policy Service to geet information related to the access rights or the users.
RIS exploits the featuress of the NOVI information model to improve the precission
of resource discovery and too apply reasoning when selecting resources and services. It
uses a database engine baseed on semantic web technologies, namely Sesame [11], for
resource selection. The datta are stored in the Sesame database as RDF triples [112].
The RIS uses the AliBaba tool [13] for the conversion of triples to Java objects. Thhese
Java objects that describe the concepts in the NOVI IM are used also by the otther
software components of thee NOVI C&M architecture.
3.3 Intelligent Resourcee Mapping (IRM) Service

The Intelligent Resource Mapping
M (IRM) service for NOVI will enable embeddding
user requests for virtual topologies
t - resources (Virtual Networks - VNs) to the
federated physical substrate network. This was initially formulated for a single
domain (infrastructure) as Virtual Network Embedding (VNE), an NP-Hard
combinatorial problem [14]. In the NOVI federated profile, VNE had to be extended
towards a multi-domain environment via graph spitting as in [15] and intelligent
selection of intra-domain mapping.
Evaluation and testing of the embedding procedure for NOVI experiments require
the appropriate representation of a VN request, formulated using the NOVI
Information Model. The IRM gathers information from the Resource Information
Service (RIS) and the Monitoring Service regarding available resources. As a first
step, user requests for VN resources are apportioned to infrastructures that are
members of a NOVI federation. Subsequently, single-domain VNE problems are
formulated, resulting into sub-optimal allocation of virtual resources within the
federated substrate.
A user may submit requests for standalone virtual resources, topologies of virtual
resources and specific services regarding virtual resources/topologies. These requests
may request specific mappings of virtual resources to substrate infrastructures. As
specified by the ProtoGENI RSpec [16]. VN requests may contain a complete (pre-
specified, bound), partial, or empty (free, unbound) mapping between virtual
resources and available physical (substrate) resources.
3.4 Policy Service

The Policy Service is used to provide the functionality of a policy-based management
system, where policies are used to define the behavior governing the managed
environment. As reported in [17], we plan to extend the Ponder2 policy framework
[18] with functions to support enforcement of mission policies. These will be used to
define the obligations of a member-infrastructure within a NOVI federation.
We currently provide support for (1) Access Control policies that specify what
rights users have on specific resources and (2) Event-Condition-Action policies
enforcing management actions upon events indicating failures or performance
degradation. Events are received by the Monitoring Service. Implementation details
are reported in NOVI Public Deliverable D2.2 [8].
3.5 Monitoring Service (MS)

One of the main challenges for a Monitoring Service (MS) in a heterogeneous
federated virtual environment is the diverse combination of monitoring tools deployed
within different infrastructures. To address this, NOVI developed generic Monitoring
Ontologies, enabling us to describe, parameterize and use diverse active and passive
monitoring tools provided within constituent federated infrastructures. Thus, users are
required to specify metrics to be measured and do not rely on monitoring tools.
MS collects information about specific resources and measurement metrics.
Monitoring can be performed on slivers (virtualized resources allocated to a user) or
on the physical substrate resources (hosts, links, paths, etc.). It is possible to obtain
passive monitoring information from resources or from repositories, and active
monitoring information as requested. Depending on usage scenario, MS can support
two main tasks: The first task is triggered by the Resource Information Service prior
to resource allocation to collect monitoring and measurement information from the

substrate, which can be used by the IRM service to ensure that the constraints defined
in the resource requests are satisfied. The second task is used after the resource
reservation, to perform slice monitoring for diagnostic and watchdog purposes, i.e. to
check the current status of a given set of virtual resources across a NOVI federation.
MS supports three advanced high level monitoring tools, i.e. SONoMA [19],
HADES [20], and Packet Tracking [21]. These tools enable users to measure key
performance metrics of the network, for example the one-way delay, the round-trip
time, the packet loss, or the available bandwidth. Obviously, the MS can obtain from
hosts via command line SSH CPU utilization, memory consumption, disk usage etc.
In Fig. 3, we provide a screenshot of the MS GUI provided to users of NOVI, who
can choose from available metrics and specify required and optional parameters. Note
that users do not need to specify which monitoring tool will measure the selected
metric, as these can vary across infrastructures (testbeds) in a NOVI federation.
Measurements of selected metrics, associated with monitoring tasks, can be managed
individually, independently from the other monitoring task. The monitoring tasks can
be started, stopped or removed from the task list. The results of the measurements can
be read from the console of the GUI, or uploaded to a database within the Resource
Information Service, or trigger event-condition-action policies in the Policy Service.
Fig. 3. GUI of the NOVI Monitoring Service
3.6 Request Handler Service

The main purpose of this service is to perform two types of operations: (1) Handling
of resource allocation requests to the underlying platforms and (2) handling external
calls coming from testbeds that are members of a NOVI federation.
For the first type, the NOVI IM needs to be translated into the underlying platform
resource specification model. Given the key role played by SFA in the federation of
PlanetLab and FEDERICA for NOVI experiments, a translation needs to be
performed between NOVI IM concepts and the ones in SFA RSpec v2 [16].
Translation in the opposite direction is needed to handle remote calls from the
federated platforms. External calls from underlying platforms occur when the Resource
Information Service (RIS) needs updates with new information, i.e. the presence of new
resources or resource status notification updates received from the Monitoring Service
(MS). RIS will only store the static part of the information from the monitoring
ontology, while the dynamic parts will be obtained by directly calling MS.
The Request Handler, as shown in Fig. 1, communicates via RSpec with a server
running SFA. Since the SFA code was initially developed for PlanetLab, we just had
to installa private SFA server for the PlanetLab part of our testbed. However, there is
no SFA implementation for FEDERICA; thus we developed an appropriate
FEDERICA RSpec and an FEDERICA SFA Wrapper service acting as FEDERICA’s
Aggregate Manager (see Fig. 1). More information can be found in D2.2: First
Information and Data Models [8].
3.7 The NOVI Distributed Virtual Switch - NSwitch

The NSwitch distributed software complements NOVI’s federation architecture by
providing a unified way of interaction between heterogeneous domains at the data-
plane. It enables a virtual entity in one domain to be connected at protocol layer 2
(L2) with another virtual entity in a remote domain taking into account concurrence,
isolation, elasticity and programmability aspects.
The NSwitch was developed, deployed and tested over the PlanetLab – FEDERICA
testbed above. Its functionality was driven by the need to combine virtual resources
belonging to these two virtualization infrastructures that employ dissimilar
communication protocols and hypervisors. Notably, PlanetLab does not provide users
with data-plane connectivity options, using IP/BGP over the legacy Internet. By
contrast, FEDERICA provides users with data-plane network virtualization choices,
e.g. providing Juniper Logical Routers and Ethernet switches based on L2 VLAN
technology. User-configurable VLANs are carried by SDH 1 Gbps circuits
provisioned by NRENs and GÉANT into a controlled WAN environment, thus
enabling repeatability of experiments over the FEDERICA infrastructure.
In order to map PlanetLab slivers into an L2 broadcast domain we adopted an
approach similar to the one developed within the VINI [22] project in the US that
introduced a set of extensions to the PlanetLab kernel and tools. VINI used an
Ethernet over GRE – EGRE [23] mechanism to provide point-to-point virtual network
capabilities to user configured virtual resources over the Internet. NOVI’s NSwitch
enhanced VINI’s capabilities by introducing the Open vSwitch (OVS) [24] S/W in
PlanetLab’s host OS, thus enabling point-to-multipoint virtual links. OVS, compared
to the VINI multiple bridges, provides better management of multiple EGRE tunnels
within a host. Furthermore, distributed OVS instances can be centrally managed via
the OpenFlow protocol [25]. This feature will be adopted in NOVI’s Spiral 2 phase.
On the FEDERICA side, L2 data plane connectivity is provided by means of
VLANs used by Logical Routers, Switches and VMs. The NSwitch functionality
provides the mapping of EGRE key values of packets originating from PlanetLab
slivers to VLAN IDs.
3.8 Integration of NOVI C&M Services

In each platform (member of a NOVI federation) the C&M Services components of
the top layer in Fig. 1 interact with each other and communicate (1) northbound with
the NOVI GUI and (2) south-bound with the middle layer (SFA). The north-bound
interface is the NOVI API, while the south-bound interface is the Request Handler
Service. Intra-domain C&M Services within the top layer exchange messages via an
Enterprise Service Bus - ESB [26]. Inter-domain C&M services can communicate (1)
via the Request Handler using SFA services (e.g. for slice creation across domains) or
(2) directly in a peer-to-peer mode via secure RPCs in cases that SFA mechanisms
were deemed as inadequate (e.g. for remote interactions of monitoring services).
An example of C&M service integration is the Slice Creation Use Case detailed in
NOVI Public Deliverable D4.2: Use Cases [27], which also provides an overview of
initial usage scenarios of the project. In summary, an authenticated experimenter is
authorized to use a set of resources across domains, as confirmed by the relevant per-
domain Policy Services. He may then request a desired topology using the NOVI GUI.
The virtual topology request is then passed to the IRM through the NOVI API. Prior
to solving the inter-domain VNE, IRM contacts RIS to identify available resources
that would fulfill the requirements imposed by the experimenter. RIS interacts with
the Monitoring Service to obtain information regarding the status (e.g. availability,
capacity, usage) of resources. Finally, when an appropriate mapping of virtual-to-
substrate resources is identified, reservation requests in the form of RSpecs are sent by
the Request Handler to the relevant testbed(s) slice manager(s).
NOVI developed a software integration framework for its C&M Services
architecture. It follows the Service Oriented Architecture complemented with the
Event Driven Architecture to enable synchronous and asynchronous communication
between components. The integration framework was implemented using Java
technologies. However it supports communication of components written in different
programming languages via a range or specific bridges such as: Jython [28], a Python
engine for Java; JRuby [29] for the Ruby language; JNI [30], a Java Native Interface
API for components written in C/C++.
4 NOVI Experimentation Testbed

To test and validate NOVI’s prototypes, a testbed environment was configured
consisting of private PlanetLab and FEDERICA resources. This testbed enables
NOVI software developers to run, test, refine and validate their software components
and prototypes, according to the experimentally driven methodology followed in the
project. NOVI developers are able to configure operational slices within the NOVI
testbed, in isolation from production services of the two virtualization platforms.
In fact, the testbed uses the actual FEDERICA substrate and virtualization services,
i.e. Juniper MX480 Logical Routers [31] and VMWare ESXi [32] Virtual Machines
(VMs). By contrast, the public PlanetLab could not be used as is for NOVI’s
experiments that require S/W upgrades, embedding custom code to C&M tools within
MyPLC [33] and root access rights to host hypervisors. Note that PlanetLab is a
widely used federated infrastructure [34], consisting of PLC (PlanetLab Central), PLE
(PlanetLab Europe) and PLJ (PlanetLab Japan), each with a single instance of
MyPLC. Experiments affecting PlanetLab’s host OS and C&M S/W are usually
performed on private testbed installations and this practice was also adopted in NOVI.
Fig. 4. Example of an experimentation slice in the NOVI testbed
Fig.4 presents the topology of one operational slice used to test control and
management plane components, detailed in NOVI Public Deliverable D4.1: Software
Architecture Strategy & Developers’ Guide [35]. This slice is comprised of three
FEDERICA core PoPs located in PSNC (Poznan, Poland), DFN (Erlangen, Germany)
and GARR (Milano, Italy). These are connected over the Internet via GRE tunnels to
private PlanetLab nodes in NTUA (Athens, Greece), ELTE (Budapest, Hungary) and
PSNC (Poznan Poland).
To isolate the slice in Fig. 4 from other NOVI slices using the same FEDERICA
core PoPs, Logical Routers are created on the Juniper MX480 routers. The open
source MyPLC (PlanetLab’s C&M software) is deployed at PSNC, managing the
private PlanetLab testbed.
An illustration of a typical slice deployed in the NOVI testbed is the NOVI-
MONITORING devoted for validating NOVI’s monitoring methods (active and
passive) and their corresponding tools. Measurements assembled via this slice are
depicted in Fig. 3.
5 Summary and Future Work

This paper reported a summary of NOVI’s current research outcomes: The ontology-
based NOVI information model, the advanced services within NOVI’s federated
Control & Management architecture and the distributed virtual switch architecture
(NSwitch). More technical details on the aforementioned work can be found in
NOVI's public deliverables and publications that are available at the project's website,
http://www.fp7-novi.eu.
NOVI’s research revealed a plethora of areas requiring further investigation. We
list below some of them:
• Information Model: Need for constant updating of NOVI’s IM evolutionary
ontologies, e.g. incorporating NSwitch parameters. Short-term ontology
enhancements based on feedback coming from validation of the integrated

prototype implementation.
• GUI: Implementation of a user feedback mechanism and support for grouping of
graphical objects, thus simplifying the level of information details of user
requests.
• Resource Information Service: Validation of the distributed architecture model
and support for more complex semantic queries that aim to provide efficient
resource discovery mechanisms, towards facilitation of virtual network
embedding processes.
• Monitoring Service: Implementation of a mechanism allowing different
monitoring tools over dissimilar platforms to cooperate and contribute to multi-
domain measurements of the same metric.
• Policy Service: Definition and deployment of role based access control (RBAC)
policies and enhancement of the policy engine to support enforcement of inter-
domain obligation Ponder2 policies
• NSwitch: Integration of the NSwitch control plane with the other components of
NOVI's C&M plane.
We are currently at the end of Spiral 1 of the project, having deployed a first version
of an integrated prototype on the NOVI testbed (Section 4 of this paper). It is
expected that the Spiral 2 subsequent effort will complement functionality and
performance of NOVI's prototypes, based on Spiral 1 results obtained from validation
experiments on the NOVI experimental testbed.
References
[1] NOVI FP7 STREP Project, http://www.fp7-novi.eu
[2] Szegedi, P., Figuerola, S., Campanella, M., Maglaris, V., Cervelló-Pastor, C.: With
Evolution for Revolution: Managing FEDERICA for Future Internet Research. IEEE
Communications Magazine 47(7), 34–39 (2009)
[3] PlanetLab, http://www.planet-lab.org
[4] GÉANT, http://www.geant.net/pages/home.aspx
[5] Global Environment for Network Innovations (GENI), http://www.geni.net/
[6] D2.1: Information Models for Virtualized Architectures, http://www.fp7-novi.
eu/index.php/deliverables
[7] Web Ontology Language (OWL), http://www.w3.org/TR/owl-features
[8] D2.2: First Information and Data Models, http://www.fp7-novi.eu/index.
php/deliverables
[9] Slice Federation Architecture, v2.0, http://groups.geni.net/geni/
attachment/wiki/SliceFedArch
[10] Ontology instance editor - OIntEd, http://novi-im.appspot.com
[11] Openrdf Sesame, http://www.openrdf.org
[12] Resource Description Framework - RDF, http://www.w3.org/RDF/
[13] Alibaba, http://www.openrdf.org/alibaba.jsp
[14] Mosharaf Kabir Chowdhury, N.M., Boutaba, R.: Network Virtualization: State of the
Art & Research Challenges. IEEE Communications Magazine 47(7), 20–26 (2009)
[15] Houidi, I., Louati, W., Ameur, W.B., Zeghlache, D.: Virtual network provisioning
across multiple substrate networks. Elsevier Computer Networks 55, 1011–1023 (2011)
[16] RSpec, http://www.protogeni.net/trac/protogeni/wiki/RSpec2rac/
protogeni/wiki/RSpec
[17] Lymberopoulos, L., Grosso, P., Papagianni, C., Kalogeras, D., Androulidakis, G., van
der Ham, J., de Laat, C., Maglaris, V.: Managing Federations of Virtualized
Infrastructures: A Semantic-Aware Policy Based Approach. In: Proc. of 3rd IEEE/IFIP
International Workshop on Management of the Future Internet, Dublin, Ireland, May 27
(2011)
[18] Ponder2, http://ponder2.net
[19] Hullár, B., Laki, S., Stéger, J., Csabai, I., Vattay, G.: SONoMA: A Service Oriented
Network Measurement Architecture. In: Korakis, T., Li, H., Tran-Gia, P., Park, H.-S.
(eds.) TridentCom 2011. LNICST, vol. 90, pp. 27–42. Springer, Heidelberg (2012)
[20] https://wiki.man.poznan.pl/perfsonar-mdm/index.php/Hades_MA
[21] Santos, T., Henke, C., Schmoll, C., Zseby, T.: Multi-hop packet tracking for
experimental facilities. In: SIGCOMM 2010, New Delhi, India, August 30-September 3
(2010)
[22] VINI, http://www.vini-veritas.net
[23] Farinacci, D., Li, T., Hanks, S., Meyer, D., Traina, P.: RFC 2784, Generic Routing
Encapsulation (GRE) (March 2000)
[24] Open vSwitch, http://openvswitch.org/
[25] OpenFlow, http://www.openflow.org
[26] Chappell, D.: Enterprise Service Bus. O’Reilly (June 2004) ISBN 0-596-00675-6
[27] D4.2: Use Cases, http://www.fp7-novi.eu/index.php/deliverables
[28] Jython, http://www.jython.org
[29] JRuby, http://jruby.org
[30] Java Native Interface, http://java.sun.com/docs/books/jni/
[31] Juniper MX480, http://www.juniper.net/us/en/local/pdf/
brochures/1500027-en.pdf
[32] VMware ESXi, http://www.vmware.com/files/pdf/ESXi_
architecture.pdf
[33] MyPLC, http://www.planet-lab.org/doc/myplc
[34] PlanetLab federation, http://www.planet-lab.org/federation
[35] D4.1: Software Architecture Strategy and Developers’ Guide, http://www.fp7-
novi.eu/index.php/deliverables
Next Generation
Flexible and Cognitive Heterogeneous Optical Networks
Supporting the Evolution to the Future Internet
Ioannis Tomkos1,*, Marianna Angelou1, Ramón J. Durán Barroso2,

Ignacio de Miguel2, Rubén M. Lorenzo Toledo2, Domenico Siracusa3, Elio Salvadori3,
Andrzej Tymecki4, Yabin Ye5, and Idelfonso Tafur Monroy6
1
Athens Information Technology (AIT), Greece
2
University of Valladolid, Spain
3
CREATE-NET, Italy
4
Telekomunikacja Polska, Poland
5
Huawei, Germany
6
Fotonik Department, DTU, Denmark
[email protected]
Abstract. Optical networking is the cornerstone of the Future Internet as it

provides the physical infrastructure of the core backbone networks. Recent
developments have enabled much better quality of service/experience for the
end users, enabled through the much higher capacities that can be supported.
Furthermore, optical networking developments facilitate the reduction of
complexity of operations at the IP layer and therefore reduce the latency of the
connections and the expenditures to deploy and operate the networks. New
research directions in optical networking promise to further advance the
capabilities of the Future Internet. In this book chapter, we highlight the latest
activities of the optical networking community and in particular what has been
the focus of EU funded research. The concepts of flexible and cognitive optical
networks are introduced and their key expected benefits are highlighted. The
overall framework envisioned for the future cognitive flexible optical networks
are introduced and recent developments are presented.
Keywords: Optical Networks, Optical Transport, Cognitive Networks, Flexible

Optical Networks.
1 Introduction
After the establishment of the first fiber-based telecom networks in the 1980s, it was
the emergence of Wavelength Division Multiplexing (WDM) a decade later that
enabled the current expansion of Internet. In these early steps of WDM networks
though, each optical channel had to be converted to the electrical domain and then
back to the optical at every node even if the optical channel was not destined for that
*
Corresponding author.
226 I. Tomkos et al.
node; these networks are commonly referred to as opaque networks. Later on, the idea
of avoiding all these costly O/E/O conversions triggered the development of Optical
Add-Drop Multiplexers (OADMs) that, in turn, allowed the establishment of
transparent networks, where the signal propagates all-optically from source to
destination throughout the network. In transparent networks, the regeneration-related
costs of opaque networks are eliminated [1] achieving up to 50% of cost saving when
compared with opaque networks [2]. Furthermore, reconfigurable OADMs
(ROADMs) and Optical Cross-Connects (OXCs) were implemented to achieve a
higher degree of flexibility and to enable networks to adapt remotely and on-demand
to the potential traffic changes, thus reducing the associated operational costs.
Moreover, the introduction of high data-rate transmission technology aims to provide
large trunks so as to accommodate the bandwidth-intensive new multimedia
applications. Nevertheless, not all traffic demands require such high bit rates and
operators are seeking for networks that are not wasting resources but are cost-
effective and therefore versatile. In this framework, existing 10 Gb/s optical networks
may upgrade their infrastructure gradually migrating to heterogeneous networks that
accommodate mixed 10/40/100 Gb/s traffic [3]. This new solution is known as Mixed
Line-Rate (MLR), as opposed to the legacy one, also referred to as Single Line-Rate
(SLR).
However, the above cited solutions provide limited flexibility and are not capable to
scale to the envisioned capacities of the Future Internet. In fact they operate under added
complexity and cost due to the rigid wavelength granularity of the systems currently
deployed. Operators provide connections with capacity that fulfils the highest (worst
case) demand (over-provisioning), while these connections remain underutilised for
most of the time. To this account, the recent advances in coherent technology, software-
defined optics and multicarrier transmission tecniques, such as Orthogonal Frequency
Division Multiplexing (OFDM) [4]-[5] and Nyquist WDM (N-WDM) [6], have
introduced the possibility to achieve a significantly high spectrum-efficiency providing
a fractional bandwidth feature. In fact, thanks to these technologies it is possible to
dynamically tune the required bit-rate and the optical reachability by appropriately
choosing the allocation of the spectrum and the modulation format. Some of the terms
often associated in literature to the optical networks exploiting these technological
advancements are “flexible”, “tunable”, “elastic” or “adaptive”. Hence, flexibility
means that the network is able to dynamically adjust the resources in an optimal and
elastic way according to the continuous varying traffic conditions, These new concepts
will enable a new network architecture where any two nodes can be connected with the
amount of bandwidth required, either providing a sub-wavelength service or super-
channel connectivity [7]-[8].
On the other side, the aforementioned emerging heterogeneous networks have
introduced a new type of challenge in network design. In reconfigurable single line-
rate networks, the resources at hand during the design phase were limited to the
channels considered feasible according to Quality of Transmission (QoT) parameters
(through physical-layer aware processes [9]) while the rate and the modulation format
were fixed. The new heterogeneous network paradigms have introduced an additional
level of flexibility, also interpreted as additional complexity. In this context, to serve a
given traffic demand, the network manager has to select the route, the channel, the
bit-rate and the modulation format [8]. Hence, traditional Routing and Wavelength
Next Generation Flexible and Cognitive Heterogeneous Optical Networks 227
Assignment (RWA) algorithms are no longer applicable and it is transformed to a

Routing, Modulation Level and Spectrum Allocation (RMLSA) problem where every
connection request is assigned a spectrum fraction.
Once the network planning has taken place, an advanced control plane solution
needs to be designed and developed in order to fully support all the aforementioned
enhancements to the optical infrastructure. Literature presents some proposals on the
control plane solutions for the physical-layer aware optical networks [9]-[11], while
the study on the flexible networks is still on a early stage, both from the
standardization (the Internet Engineering Task Force has recently published some
internet drafts [12]-[13]) and the research point of view (very few works have been
published, among them [14]). Through properly developed Generalized Multi-
Protocol Label Switching (GMPLS) protocol extensions, the control plane is expected
to be able to support the overall networking solution and allows the different building
blocks to cooperate and run in an orchestrated manner. On the whole, the concepts of
physical-layer awareness and of spectrum flexibility will require intelligent
techniques to offer optimal static planning, dynamic configuration and management
of optical signal with acceptable QoT. In such a context, the control and management
planes will work in conjunction to provide dynamic routing and flexible spectrum
assignment, management of sub-wavelength service or super-channel connectivity,
performance and impairment monitoring, traffic monitoring, failure localization, etc.
In turn, the information stemming from the data plane considered valuable to the
various modules will be disseminated to the nodes of the network through properly
enhanced control plane extensions.
A promising solution to tackle these challenges comes from exploiting cognition
[15]. The use of cognitive techniques in optical networks brings about an extended
level of “intelligence” to the optical layer by facilitating the adaptive tuning of various
physical layer characteristics (modulation format, forward error correction,
wavelength capacity, etc) and network layer parameters (bandwidth, number of
simultaneous lightpaths, QoS, etc) depending on application or service requirements.
Cognitive networks typically perform cross-layer design and multi-objective
optimization in order to support trade-offs between multiple goals; thus they become
a promising option to optimize the performance of optical networks in a cost- and
energy-efficient way. This approach is fully aligned with a Future Internet vision
where the role of an optical network is not just about providing fixed high-speed
bandwidth between node pairs, but instead it enables network operators to finely tune
these pipes among a set of nodes in order to provide an “application-specific” virtual
network which complies delay and bandwidth constraints according to application
requirements. As envisioned in projects more focused on Future Internet services like
the ones currently pursued within the FI-PPP framework [16], these requirements can
in fact be wildly different and thus deserve a highly adaptive transport layer. A
cognitive optical network based on flexible-grid technologies will strengthen the link
between the client (IP) and the transport (optical) layer [17] to an extreme degree thus
providing a consistent contribution toward the EU vision of a future network
infrastructure that support the convergence of heterogeneous broadband technologies
as enablers of the Future Internet.
In this chapter, we present the approach followed within the FP7 European
Cognitive Heterogeneous Reconfigurable Optical Network (CHRON) project [18], in
which a cognitive architecture is proposed in order to realise a flexible optical Future

Internet infrastructure. The investigated cognitive solution is expected to provide
effective multilayer decisions on (i) how to efficiently route traffic over the network;
and (ii) how to allocate the spectrum and choose the appropriate
transmission/switching technique, optical launch power, modulation format, bit-rate,
etc., thus relying in cross-layer design techniques. On the other hand, we also
demonstrate the advantages of using heterogeneous flexible networks in terms of
three parameters: the spectrum efficiency, the cost and the energy consumption.
2 Cognitive Optical Networking

A cognitive network is defined as “a network with a process that can perceive current
network conditions, and then plan, decide, and act on those conditions. The network
can learn from these adaptations and use them to make future decisions, all while
taking into account end-to-end goals” [15]. Therefore, a cognitive network should
provide better end-to-end performance than a non-cognitive network. In fact,
cognition has already been tested and proven to be an excellent solution for wireless
networks [19].
However, cognitive networks are also applicable to wired communication
architectures, and are especially appealing for optimizing performance in
heterogeneous networks. Since cognitive networks typically perform cross-layer
design and multi-objective optimization in order to support trade-offs between
multiple goals, they also become a promising option to optimize the performance of
heterogeneous optical networks in a cost efficient way.
In the area of optical communications, cognitive techniques are exploited in the
framework of CHRON [18] project so to enable “intelligence” in the optical layer. In
particular, CHRON should be able to provide effective decisions, by relying on
cognition, on:
• how to route new traffic demands, either through existing optical connections
(lightpaths), through new lightpaths or by triggering a reconfiguration process of
the virtual topology (i.e., by rearranging existing connections);
• how to assign resources, not only wavelengths or spectrum, but also the most
appropriate transmission/switching technique, modulation format, bit-rate, etc.;
• how to ensure energy-efficient operation;
and all while taking into account the Quality of Service (QoS) and Quality of
Transmission (QoT) requirements of the demands.
According to the definition of cognitive networks given above, those decisions must
be made by taking into account current status and knowledge acquired through previous
experience. Thus, the core element of the CHRON architecture is the cognitive decision
system. Such a system is complemented with a network monitoring system, which
provides traffic status and optical quality of transmission measurements, and with a set
of control and management mechanisms to implement the decisions that are made by
the cognitive decision system and to disseminate the monitored information. The
interaction of those building elements is detailed in Fig. 1.
Since the cognitive decision system must deal with very diverse tasks, it is
composed by five different modules, all of them exploiting cognition. Thus, it
includes a RWA/RMLSA module to process optical connection (lightpath) requests; a
QoT estimator module to predict the QoT of the optical connections before being
established (and thus helping the RWA/RMLSA module to ensure that quality
requirements are met); a virtual topology design module, which determines the
optimal set of lightpaths that should be established on the network to deal with a
given traffic demand, and a traffic grooming module, which is in charge of routing
traffic through the lightpaths composing the virtual topology. Last but not least, a
network planner and decision maker module coordinates and triggers the operation of
the other modules and handles the communications with other network elements.
Fig. 1. Main elements of the CHRON approach
In the framework of this architecture, the advantages of cognition have already been
demonstrated in a number of scenarios, such as on quickly and effectively assessing
whether an optical connection (i.e., a lightpath) satisfies QoT requirements [20], or on
determining which set of connections should be established on an optical network (i.e.,
the so-called virtual topology) in order to support the traffic load while satisfying QoT
requirements and minimizing energy consumption and congestion [21].
In the former scenario, the utilization of Case-Based Reasoning techniques to
exploit knowledge acquired through previous experiences leads to obtaining not only
a high percentage of successful classification of lightpaths into high or low QoT
categories (Fig. 2), but also to a great reduction in the computing time (around three
orders of magnitude) when compared to a previous tool for QoT assessment which
does not employ cognition [20].
In the latter scenario, the inclusion of cognition in a multi-objective algorithm to
determine the optimal set of virtual topologies with different trade-offs in terms of
throughput and energy consumption brings great advantages. Since a multi-objective
algorithm provides a set of solutions (i.e., virtual topologies) in a single execution, we

have joined the solutions provided by two versions of the same algorithm: one
without cognition and the other with cognition. Then, the best set of solutions has
been selected, which is called the common Pareto Optimal Set (POS). Fig. 3 shows
that at the beginning (when there is no previous history that the cognitive method can
exploit), both methods contribute approximately with the same number of solutions.
However, once cognition really enters into play, i.e., when enough past history is
used, most of the solutions contained in the common POS (i.e., the best solutions) are
obtained by the cognitive method [21].
100.0 100
Successful classification of lightpaths (%)
99.9 8 wavelengths
99.8 16 wavelengths
32 wavelengths 80
99.7
64 wavelengths
% Solutions that belongs

99.6
to the Common POS

99.5
60
99.4
Cognitive Algorithm
99.3 Non - Cognitive Algorithm
99.2 40
99.1
99.0
98.9 20
98.8
98.7
98.6
0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000 0 6 12 18 24 30 36 42 48
Number of cases in the Knowledge Base Time (h)
Fig. 2. Successful classification of lightpaths Fig. 3. Percentage of solutions in the common

into high/low QoT categories POS found by a method with cognition and the
same method without cognition
3 Advantages of Mixed Line-Rate and Flexible Networks

Cognition is a useful tool capable of optimizing the design and control of an optical
network. A cognitive network allows the introduction of a flexible transport to
support the Future Internet, by pushing down to the optical layer some of the
“intelligence” typically performed in the IP layer. However, a key element for
operators seeking to migrate to the next-generation core is the evaluation of the trade-
off between the extra capital investment that it requires and its performance.
Moreover, in addition to the capital cost of the future core network, power
consumption is another parameter that becomes relevant, mainly due to the
operational economic implications, considering the pace at which traffic is increasing
annually. The goal of this section is to discuss the new mixed line-rate and flexible
core networks from a cost, spectral and energy perspective and give a comprehensive
view of the potential of each solution. Focusing on the importance of spectrum as a
resource, novel RMLSA algorithms for path and resource allocation in flex-grid
networks are exploited herein [22].
Nevertheless, to realize the level of flexibility of the multi-carrier solutions, new
network and transmission elements need to be introduced in the optical transport,
implying extra capital investment. Software-defined transponders [23] and
bandwidth-flexible optical nodes [24] employing spectrum-flexible Wavelength
Selective Switches (SF-WSS) are the key enablers for the implementation of this
architecture. The methodology presented in [25] is used to investigate the

requirements in capital of the flex-grid networks over the fixed-grid solutions in
correlation with the gained spectrum optimization. Following the optimized resource
allocation, all solutions are evaluated under the prism of energy efficiency. The
energy efficiency that each solution incurs is estimated considering the power
consumption needs of the associated networking elements.
3.1 Spectrum Allocation Advantages

The analysis considers networking solutions that can deliver up to 400 Gb/s per
channel in a fixed or flexible spectrum grid utilizing physical-layer aware algorithms
to route and allocate the available spectrum [26], [4]. The study includes fixed WDM
SLR networks that deliver either 40 Gb/s, 100 Gb/s or 400 Gb/s per channel and MLR
[9] networks with data rates of 10 Gb/s, 40 Gb/s, 100 Gb/s and 400 Gb/s. Regarding
the flex-grid solutions, two multi-carrier solutions have been considered; one refers to
the technique reported in [4] (denoted as E-OFDM) while the other refers to
the technique in [5] (denoted as O-OFDM). Both multi-carrier solutions can adapt the
transmitted bit-rate from 10Gb/s-400Gb/s by modulating subcarriers with the
necessary modulation level.
4000
SLR 40G
3500 SLR 100G
Spectrum Utilized (GHz)
SLR 400G
3000 MLR
E-OFDM
2500
O-OFDM
2000
1500
1000
500
0
1 2 3 4 5 6
Traffic Load Multiplier
Fig. 4. Spectrum utilization for all solutions and different traffic loads
To calculate the bandwidth utilized by the various solutions the Deutsche Telekom
core network (14 nodes, 23 bidirectional links) and the realistic traffic matrix of the
DT network for 2010 scaled up to 11 times to obtain traffic ranging from 3.6 Tb/s up
to 39.6 Tb/s has been utilized. Under the given assumptions, the flexible multi-carrier
solutions offer the most efficient spectrum allocation as expected from the optimized
packing of the connections in the frequency domain (Fig. 4).
3.2 Cost Efficiency Advantages

Spectrum utilization is not only used as a way to evaluate the networking solutions
but also in the form of spectrum savings (considered here in 50GHz slots) that can be
utilized for the provisioning of new traffic. Based on the methodology introduced in
[25] the total cost of a system is modeled considering three cost parameters: the cost
of transponders, the cost of node equipment and the one related to the number of
“dark” 50GHz channel slots that are utilized and associated only with the link
infrastructure cost.
Among the fixed-grid networks the distinctive component that determines the
capital requirements is the type of the transponders. Fig. 5 illustrates the absolute
number of transponders per networking solution. Fig. 6. shows the relative
transponder cost of all fixed-grid solutions; the relative cost values are set at
1/2.5/3.75/5.5 for the 10 Gb/s, 40 Gb/s, 100 Gb/s and 400 Gb/s transponders
respectively [27]. For MLR systems, two variations of the planning algorithm are
reported; the first one seeks to minimize the number of utilized wavelengths, and the
second one optimizes the transponder cost of the network.
1200 SLR 40G 5000

Transponders in absolute numbers
SLR 100G SLR 40G

4500 SLR 100G
SLR 400G
Relative Transponder Cost

1000 MLR-Optimizing TR cost 4000 SLR 400G
MLR-Optimizing wavelengths
3500 MLR - Optimizing TR cost
800 E-OFDM
O-OFDM MLR - Optimizing wavelengths used
3000
600 2500
2000
400
1500
1000
200
500
0 0
1 3 5 7 9 11 1 3 5 7 9 11
Traffic Load Multiplier Traffic Load Multiplier
Fig. 5. Required number of transponders for Fig. 6. Relative transponder cost for the
all solutions to serve the different traffic fixed-grid networking solutions
matrices (in absolute numbers)
600%
Additional cost of E-OFDM transponder
Load 1 Load 3
Additional cost of O-OFDM transponder
300%
Load 1 Load 3
Load 5 Load 7
over the 100G transponder cost
Load 5 Load 7
over the 100G transponder cost
Load 9 Load 11
500% Load 9 Load 11 250%
400% 200%
300% 150%
200% 100%
100% 50%
0% 0%
10000 25000 40000 55000 70000 85000 100000 10000 25000 40000 55000 70000 85000 100000
Cost per 50GHz channel slot Cost per 50GHz channel slot
Fig. 7. Allowable additional cost for E-OFDM Fig. 8. Allowable additional cost for O-
transponder compared to SLR 100G from OFDM transponder compared to SLR 100G
spectrum savings for different traffic loads from spectrum savings for different traffic
loads
However, reliable data for the cost of the flex-grid networks components, i.e., the
software-defined transponders and bandwidth-variable nodes, are currently not
available. To overcome this, the extra cost of the E-OFDM and O-OFDM
transponders over the cost of a 100 Gb/s transponder so as to achieve total network
cost equal to that of the related SLR network is examined. The comparison was
focused on the cost of the E-OFDM and O-OFDM transponders as those rely on
electronics for DSP. Fig. 7 presents the allowable additional cost for the E-OFDM
transponder compared to the SLR 100 Gb/s transponder for different traffic loads. For
a 50 GHz-channel cost that ranges from 10 k€ to 100 k€, an E-OFDM transponder
may cost 3 to 5 times more when the traffic load is equal to 11 so as to achieve total
network cost equal to that of the SLR network. For the lowest traffic scenario
(load=1), where the spectrum savings of the flex-grid solution compared to the 100G
SLR are less pronounced, the E-OFDM solution is preferable over the SLR network
when the additional cost that is tolerable ranges between 6% to 50%. In a similar
manner, Fig. 8 presents the results for the comparison between O-OFDM and 100G
SLR. The O-OFDM transponder may cost approximately 2-3 times more for the
highest traffic load scenario. The difference with the O-OFDM case is justified by its
higher spectrum utilization as shown in Fig. 4. From the operators’ perspective, these
results indicate how the spectrum savings of the flex-grid networks can be used to
mitigate the additional cost of the new spectrum flexible transponders.
3.3 Energy Efficiency Advantages

In addition to the capital cost of the future core network, power consumption is
another parameter that becomes relevant in network planning, mainly due to the
operational economic implications but also the growing ecological awareness,
considering the pace at which traffic is increasing annually. Following the resource
allocation of all solutions, the energy efficiency is estimated considering the power
consumption needs of the associated networking elements. Hence, the considered
solutions were compared with respect to the power consumption of the associated
network elements, i.e., transponders, optical cross-connects (OXCs) and optical line
amplifiers.
0,20 SLR 40G

0,18
Energy Efficiency (Gb/s/W)
SLR 100G
0,16
0,14 SLR 400G
0,12
MLR-Optimizing
0,10
TR cost
0,08 MLR-Optimizing
0,06 wavelengths
E-OFDM
0,04
0,02 O-OFDM
0,00
1 3 5 7 9 11
Traffic Load Multiplier
Fig. 9. Energy Efficiency achieved for all solutions and different traffic loads
The estimated energy efficiency (in Gb/s/W) for the various traffic loads is
illustrated in Fig. 9. 400G SLR appears to be the least efficient for traffic load up to 5
although it tends to improve for higher loads. The other SLR solutions achieve better
efficiency that decreases for high loads justified by the great number of transponders
as depicted in Fig. 5. On the other hand, the granularity of 10G/40G/100G/400G in

MLR and of the low-rate subcarriers in O-OFDM appears to be sufficient for the
entire range of traffic loads optimizing the number and type of transponders and
leading to low power consumption. Under the given power consumption assumptions,
E-OFDM demonstrates lower energy efficiency for load up to 5. Moving up in traffic
load, the transponders assumed run at higher bit rates leading to superior energy
efficiency.
On the whole, in terms of the overall network energy efficiency, flex-grid solutions
achieve low energy per bit as they use just the amount of network resources needed
for given input traffic.
4 Conclusions
Optical networking developments allow the reduction of complex operations at the IP
layer so as to reduce the latency of the connections and the expenditures to deploy and
operate the networks. New research advancements in optical networking promise to
further fortify the capabilities of the Future Internet. In this context, the CHRON
project proposes a Cognitive Heterogeneous Reconfigurable Optical Network, which
observes, acts, learns and optimizes its performance, taking into account its high
degree of heterogeneity with respect to QoS, transmission and switching techniques.
The aim of CHRON is to develop and showcase a network architecture and a control
plane which efficiently use resources in order to minimize CAPEX and OPEX while
fulfilling QoS requirements of each type of service and application supported by the
network in terms of bandwidth, delay and quality of transmission, and reducing
energy consumption.
The cognitive process and the consequent cross-layer proposed solutions have been
extensively exploited to deliver connections at a single line-rate. Nevertheless due to
their potential, flexible optical networking solutions have been investigated within the
CHRON project, as well as their predecessor, the mixed line-rate (MLR) one. In order
to demonstrate the potential of cognitive techniques, we have shown the performance
advantages brought when cognition is used in two different scenarios: the estimation
of the QoT of the lightpaths established (or to be established) in an optical network,
and the design of efficient virtual topologies in terms of throughput and energy
consumption. Then, the advantages of flexible optical networks have been evaluated.
As opposed to the rate-specific and fixed-grid solution of an MLR network,
flexible optical networks, regardless of the employed technology, are bandwidth
agnostic and have the ability to deliver adaptive bit-rates. The associated technologies
and concepts that enable the vision of flexible optical networks include advanced
modulation formats that offer higher spectral efficiency, the concept of a spectrum-
flexible grid, software-defined optical transmission, single-carrier adaptive solutions
and multi-carrier technologies. Nevertheless the increased level of flexibility imposes
complex requirements with respect to the spectrum and capacity allocation.
Therefore, in this context, CHRON has evaluated the core networks of the Future
Internet from a cost, spectral and energy perspective and has provided a
comprehensive view of the potential of various technologies. This investigation has
been carried out by taking into account the greatly different requirements of Future
Internet application as well as the need for energy-efficient future network

infrastructures that support the convergence and interoperability of heterogeneous
mobile, wired and wireless technologies, as envisioned in the EU FP7 research
framework. The resource optimization achieved in MLR and flexible networks has
been investigated under the prism of cost and energy efficiency. First a methodology
has been introduced to explore the conditions under which the vision of flexible
networking makes a good business case. Single and multi-carrier networks offering
channel rates up to 400 Gb/s have been evaluated under realistic reach parameters.
The aforementioned methodology has been applied to examine how the efficient
spectrum utilization and fine bit-rate granularity of flex-grid core optical networks
may affect the requirements in capital and power compared to fixed-grid solutions. It
has been shown that the capability of the flex-grid networks to allocate efficiently the
available spectrum counterbalances the additional capital expenditures that are
required to migrate to a multi-carrier system. On the whole, in terms of the overall
network energy efficiency, flex-grid solutions can achieve low energy per bit as they
use just the amount of network resources needed for the given input traffic.
References
[1] Saleh, A.A.M.: Transparent optical networking in backbone networks. In: Optical Fiber
Communication Conference, OSA Technical Digest Series Optical Society of America
(2000); paper ThD7
[2] Gunkel, M., et al.: A Cost Model for the WDM Layer. In: Photonics in Switching
Conference PS (2006)
[3] Liu, X., Chandrasekhar, S.: High Spectral-Efficiency Mixed 10G/40G/100G
Transmission. In: Proc. AOE 2008, OSA Technical Digest (CD) (Optical Society of
America) (2008); paper SuA2
[4] Klekamp, A., et al.: Transmission Reach of Optical-OFDM Superchannels with 10-600
Gb/s for Transparent Bit-Rate Adaptive Networks. In: Proceedings of ECOC (2011);
paper Tu.3.K.2
[5] Chandrasekhar, S., et al.: Transmission of a 1.2-Tb/s 24-Carrier No-Guard-Interval
Coherent OFDM Superchannel over 7200-km of Ultra-Large-Area Fiber. In:
Proceedings of ECOC (2009); paper PD2.6
[6] Gavioli, G., et al.: Investigation of the Impact of Ultra-Narrow Carrier Spacing on the
Transmission of a 10-Carrier 1Tb/s Superchannel. In: Proceedings of OFC (2010);
paper OThD3
[7] Borkowski, R., et al.: Experimental Demonstration of Mixed Formats and Bit Rates
Signal Allocation for Spectrum-flexible Optical Networking. In: Proc. Optical Fibre
Communications Conference, Los Angeles CA, USA (March 2012); paper OW3A.7
[8] Jinno, M., et al.: Spectrum-efficient and scalable elastic optical path network:
architecture, benefits, and enabling technologies. IEEE Communications Magazine 47,
66–73 (2009)
[9] Agraz, F., et al.: Experimental Demonstration of Centralized and Distributed
Impairment-Aware Control Plane Schemes for Dynamic Transparent Optical Networks.
In: OFC/NFOEC (2010); paper PDPD5
[10] Nag, A., Tornatore, M., Mukherjee, B.: Optical Network Design With Mixed Line
Rates and Multiple Modulation Formats. IEEE/OSA Journal of Lightwave Technology
(JLT) 28, 466 (2010)
[11] Salvadori, E., et al.: Handling Transmission Impairments in WDM Optical Networks by
Using Distributed Optical Control Plane Architectures. IEEE/OSA Journal of
Lightwave Technology (JLT) 27(13) (July 2009)
[12] Zhang, F., et al.: Requirements for GMPLS Control of Flexible Grids. IETF Internet
Draft. draft-zhang-ccamp-flexible-grid-requirements-01.txt (October 2011)
[13] King, D., et al.: Generalized Labels for the Flexi-Grid in Lambda-Switch-Capable
(LSC) Label Switching Routers. IETF Internet Draft, draft-farrkingel-ccamp-flexigrid-
lambda-label-01.txt (October 2011)
[14] Cugini, F., et al.: Demonstration of Flexible Optical Network based on Path
Computation Element. IEEE/OSA Journal of Lightwave Technology (JLT) 30(5)
(December 2011)
[15] Thomas, R.W., Friend, D.H., DaSilva, L.A., MacKenzie, A.B.: Cognitive networks:
adaptation and learning to achieve end-to-end performance objectives. IEEE
Communications Magazine 44, 51–57 (2006)
[16] Future Internet Public-Private Partnership Programme (FI-PPP), http://ww.
fi-ppp.eu
[17] Gerstel, O.: Realistic Approaches to Scaling the IP Network using Optics. In:
Proceedings of OFC (2011); paper OWP1
[18] CHRON project, http://www.ict-chron.eu
[19] Mahmoud, Q.H. (ed.): Cognitive Networks: Towards Self-Aware Networks. John
Wiley & Sons, Ltd. (2007)
[20] Jiménez, T., et al.: A Cognitive System for Fast Quality of Transmission Estimation in
Core Optical Network. In: Proc. OFC/NFOEC (2012); paper OW3A.5
[21] Fernández, N., et al.: Cognition to Design Energetically Efficient and Impairment
Aware Virtual Topologies for Optical Networks. In: 16th International Conference on
Optical Networking Design and Modeling, ONDM 2012. University of Essex,
Colchester (in press, 2012)
[22] Christodoulopoulos, K., et al.: Elastic Bandwidth Allocation in Flexible OFDM-based
Optical Networks. J. Lightwave Technol. 29, 1354–1366 (2011)
[23] Bocoi, M., et al.: Cost Comparison of Networks Using Traditional 10&40 Gb/s
Transponders versus OFDM Transponders. In: Proceedings of OFC (2008); OThB4
[24] Poole, S., et al.: Bandwidth-flexible ROADMs as Network Elements. In: Proceedings
of OFC (2011); paper OTuE1
[25] Christodoulopoulos, K., et al.: Value analysis methodology for flexible optical
networks. In: ECOC 2011 (2011); paper We.10.P1.89
[26] Christodoulopoulos, K., Manousakis, K., Varvarigos, E.: Reach Adapting Algorithms
for Mixed Line Rate WDM Transport Networks. J. Lightwave Technol. 29, 3350–3363
(2011)
[27] Patel, A.N., Ji, P., Jue, J.P., Wang, T.: First Shared Path Protection Scheme for
Generalized Network Connectivity in Gridless Optical WDM Networks. In:
Proceedings of ACP 2010, PD6, 1-2 (December 2010)
A Tentative Design of a Future Internet Networking
Domain Landscape
Marc Pallot, Brigitte Trousse, and Bernard Senach
INRIA Sophia Antipolis – Méditerranée, AxIS Project-team, France
Abstract. The Future Internet (FI) will dramatically broaden both the spectrum
of available information and the user’s possible contexts and situations. This
will lead to the vital need of a more efficient use of the Internet resources for
the benefit of all. While the Internet has already delivered huge economic and
social benefits over its short lifespan, there must be a realignment of how
Internet research and investments are made and value is captured for enabling a
continuous growth. The increase of available online contents and networking
complexity require the exploration, experimentation and evaluation of new
performance optimisation approaches for delivering different types of contents
to users within different contexts and situations. Several network research areas,
such as peer-to-peer, autonomous, cognitive and ad hoc networking, have
already demonstrated how to improve network performance and user
experience.
Interestingly, there are various Internet-networking research areas and
corresponding technologies that were investigated, experimented and
progressively deployed, while others emerged more recently. However, there
are still open questions such as visualising the conceptual evolution and
articulating the various FI networking and computing research areas and
identifying appropriate concepts populating such a FI domain landscape. This
paper presents a tentative FI domain landscape populated by Internet computing
and networking research areas.
Keywords: Future Internet, Internet-Networking, Domain-Landscape, Network-

Computing, Internet-Routing.
1 Introduction
The Internet has progressively become a ubiquitous environment for globally
communicating and disseminating information. There is a limitless amount of
available online resources and tools to share information and develop a better
understanding on whatever topics. With the recent advent of user created content,
thanks to the web 2.0 social approach, there has been a tremendous expansion in the
number of web pages created every day for exposing and sharing societal issues such
as environmental monitoring, energy efficiency, food and drug security as well as
human well-being. Tools like photo/video sharing, mash-ups, tagging, wikis and
collaborative virtual worlds enable new ways for the society to explore and
understand past present and future challenges. The Future Internet (FI) will
238 M. Pallot, B. Trousse, and B. Senach
dramatically broaden both the spectrum of available information and the user’s
possible contexts and situations. This will lead to the vital need of a more efficient use
of the Internet resources for the benefit of all. While the Internet has already delivered
huge economic and social benefits over its short lifespan, there must be a realignment
of how Internet research and investments are made and value is captured for enabling
a continuous growth.
2 Future Internet Networking Domain Landscaping
2.1 The Future Internet

Pirolli and colleagues [1] argue that an extensive research is required for building
upon currently used Internet media and tools to foster wider user participation to
tackle US national priorities through technology-mediated social participation. Mobile
and Internet technologies have converged into ubiquitous social connectivity. Pirolli
and colleagues report that in spring 2010, 40% of adults (aged 30 and over), 72% of
young adults and 73% of teens use social network web-sites and that time spent on
Facebook increased by more than 500%. They also mention that many vibrant
communities have emerged on the Web such as Wikipedia having more than 12
million registered users and more than 3 million content pages and in February 2010,
Twitter users generated 35 million tweets per day.
The term “Future Internet” (FI) represents worldwide research activities for re-
inventing the Internet with better performance, reliability, scalability, security and
privacy while keeping its key neutral principle as constantly recommended by Tim
Beemer’s-Lee, the famous inventor of the Web. As shown in the FI networking
domain landscape (see Figure 1), there is a great diversity of research streams and
related topics for designing alternatives of the Internet networking of tomorrow. For
example, the Internet of Things (IoT) is considered as a major research and
innovation stream leading to create plenty of service opportunities in interconnecting
physical and virtual worlds with a huge amount of electronic devices (e.g. sensors,
actuators) distributed in houses, vehicles, streets, buildings and many other public
environments (e.g. airports, train, metro and bus stations, social spaces). Hence, a
massive amount of data will be flowing over the Internet that should not decrease the
overall service performance and user satisfaction.
The movement towards the Future Internet is based on the belief that the current
Internet has reached his limits. Tselentis [2] states: “The current Internet has been
founded on a basic architectural premise, that is: a simple network service can be
used as a universal means to interconnect intelligent end systems. This simple premise
has allowed the Internet to reach an impressive scale in terms of inter-connected
devices. However, while the scale has not yet reached its limits, the growth of
functionality and the growth of size have both slowed down. It is now a common belief
that the current Internet would reach soon both its architectural capability limits and
its capacity limits.”
The FI represents the evolving need for infrastructures at the level of innovation
infrastructure (networks of collaboration, experimental facilities, research and test
centres etc), and broadband Internet infrastructure (networks, services). Recently,
A Tentative Design of a Future Internet Networking Domain Landscape 239
several testbeds were initiated, such as PlanetLab [3], TEFIS, BonFIRE and SensLAB
[2]. TEFIS supports the Future Internet of Services Research by offering a single
access point to different testing and experimental facilities for communities of
software and business developers to test, experiment, and collaboratively elaborate
knowledge [4], [5]. The main goal of the BonFIRE project is to design, build and
operate a multi-site Cloud prototype FIRE facility to support research across
applications, services and systems at all stages of the R&D lifecycle, targeting the
services research community on Future Internet. The purpose of the SensLAB project
is to deploy a very large scale open wireless sensor network platform, in order to
provide an accurate and efficient scientific tool to help in the design and development
of real large-scale sensor network applications. SensLAB has been instrumental in
detecting overlapping communities in complex networks [6].
Challenging issues arise from the study of dynamic networks like the
measurement, analysis and modelling of social interactions, capturing physical
proximity and social interaction by means of a wireless network. A concrete case
study exhibited the deployment of a wireless sensor network applied to the
measurement of Health Care Workers' exposure to tuberculosis infected patients in a
service unit of the Bichat-Claude Bernard hospital in Paris, France [7]. As described
above through different testbed projects, the Future Internet is the “provider” of future
Internet infrastructure and applications. Obviously, the Future Internet will be the key
driver of technological support for services and products to be explored, experimented
and evaluated.
2.2 Towards a Tentative Future Internet Networking Research Domain

Landscape
While working on the development of a Living Lab research domain landscape on
methods for involving users in R&D [8] and discussing about the scientific program
of the FIA event, it came to our mind that it could be useful to prepare a map as a
tentative FI landscape populated by Internet computing and networking research
areas. Furthermore, several INRIA research teams are involved in FIRE Testbed
projects, namely: PlanetLab, OneLab, TEFIS, SensLAB, and BonFIRE whose
scientific leaders were interviewed during the development of this tentative FI
landscape.
We believe that the proposed landscape of FI networking research domain could
provide a faster and broader understanding of the different research streams and
related topics. Several dimensions were used for landscaping the FI networking
research domain:
• Evolution approaches: from incremental (evolution) design to Clean Slate re-

design or radical evolution from where emerge new generation networks;
• Internet routing: from the basic data packet delivery towards more sophisticated
content distribution and retrieval capacities (content Centric Networking);
• Network Types: from wired communication (cable or optical networks) to
wireless communication networks (wireless Internet, wireless sensors networks);
• Evolution trends: from a traditional computer network towards an autonomic and
convergent network that become the computing network.
A number of concepts representing various research areas were selected for

populating the domain landscape. The selection of concepts was carried out in parsing
a large set of published papers in order to identify prominent terms that correspond to
research publication streams, which were validated during the interviews. The
following six categories were identified:
• Network Computing: Pervasive Computing, Ubiquitous Computing, Grid

Computing, and Cloud Computing;
• Internet Routing: Peer-to-Peer Network, Ad-hoc Network, Content Centric
Networking, Self-adaptive Network, Resilient Network, Fault tolerant Network,
Autonomic Network and Cognitive Network;
• Network Type: Wireless Sensor Network, Optical Network, Wireless Internet;
• Network Security: Virtual-private Network, Internet Security;
• Network Assessment: Quality of Services, Quality of Experience;
• Network (IP) Globalisation: Next Generation Network, IP Multimedia
Subsystem, Internet of Things, Network Convergence.
Two categories are clearly identical to two of the above mentioned dimensions,
namely: “Internet Routing” and “Network Type”. A third category, “Network
Computing”, quite overlaps with the dimension named “Evolution trends”.
As for the category “Network Computing”, it is worth to note that the concept of
Pervasive Computing, often mentioned as the ‘disappearing computing’, and
Ubiquitous Computing, rather evoked as ‘computing is everywhere’ are often used
synonymously especially in the Ambiance Intelligence area. In the same vein, the
concept of Grid computing, known as a cluster of networked computers, and Cloud
Computing, computing as a service or storage as a service, are quite closely related
from the perspective of shared resources. Regarding the category of Network
Globalisation, all the concepts are related to the convergence towards ‘all IP1’ strategy
and to the concepts of the Network Computing category as well as the Internet
Routing category. The Network Security and Network Assessment categories have
more transversal concepts that need to be considered at the earlier stage of the FI
design.
For each research stream, a Google scholar search over three different time periods
was carried out as a publication metric intended to show their respective growth or
decline. All selected concepts, considered as research areas, are individually described
in the Table 4 Appendix at the following URL2. The respective levels of publication
for each concept are provided in the table below (see Table 1) showing the
publication values for the three respective time periods, and sorted by ascending value
of the column 2006-2011.
1
Internet Protocol.
2
http://www.mosaic-network.org/pub/bscw.cgi/0/69097
Table 1. Foreseen Concepts Belonging to the FI Networking Research Domain Landscape
Concepts Number of Papers

(research areas) 2006 -2011 2000 -2005 1990 -1999
Content-centric Networking 81 1 0
Self-adaptive network 88 70 26
Resilient Network 424 179 57
Fault tolerant network 544 326 390
Autonomic Network 715 151 25
Cognitive Network 1370 377 273
Network Convergence 1760 988 566
Quality of Experience 2230 672 91
Internet of Things 2400 117 8
Optical Networking 2500 2450 234
IP Multimedia Subsystem 2800 604 3
Next Generation Network 4030 1650 206
Peer-to-Peer Network 6780 3630 122
Quality of Services 6970 5300 1050
Internet Security 7030 6130 1390
Wireless Sensor Network 7320 1140 16
Semantic Web Services 7990 2390 11
Virtual Private Network 8100 6930 844
Cloud Computing 10200 127 144
Wireless Internet 12400 8440 332
Ad hoc Network 12500 7160 291
Grid Computing 15100 7870 75
Ubiquitous Computing 15300 12200 1230
Pervasive Computing 15600 8970 129
The bar-graph below (see Figure 1) shows the growth in terms of published papers for
the respective selected concepts across the three different time-periods. The highest
level of publication belongs to the concepts of the category “Network Computing”
and Ad hoc Network as well as Wireless Internet. However, the growth rate of Cloud
Computing looks so impressive that it is quite easy to predict it as the next big thing
on the Future Internet. Not surprisingly, among other concepts having an impressive
growth rate are Wireless Sensor Network and Internet of Things. The lowest level of
published papers appears to be related to more emerging concepts of the Internet
Routing category, such as Content Centric Networking, Self-adaptive Network,
Resilient Network, Fault-tolerant Network and Cognitive Network.
The growth rate of Virtual Private Network is impressively decelerating in the last
time-period while it had an impressive growth rate in the middle time-period. The
same evidence appears to apply on Internet Security and Quality of Services. The
situation is even worst in terms of growth rate for Optical Networking, which seems
to have reached its maximum amount of annual publication.
Fig. 1. Publication Metric for FI Networking Research Concepts
The landscape is divided twice. First of all, it is divided in two spaces: a top space
and a bottom space that respectively address the wired and wireless Internet.
Secondly, it is divided in a right hand located space corresponding to the more
traditional “Computer Network” and in a left hand space representing the more recent
“Network Computing”.
A tentative design of the Future Internet networking research domain landscape for
three successive time periods appears below (see Figures 2, 3 and 4) where each
concept, presumed research area, appears as a bubble whose size is proportional to the
overall amount of publication in the corresponding time-period. The various concepts
and their allocated bubbles populate the landscape according to the four different
dimensions.
FI Networking Domain Landscape for the time period 1990-1999

The FI networking research domain landscape for the time period 1990-1999 appears
in Figure 2 where the concept of “Network Computer” forms a big island on the left
hand side due to its publication level of 16400 published papers. The opposite island,
about 5 times smaller, is constituted by the concept of “Network Computing” with
3340 published papers in the same period. The concept of “Ubiquitous Computing”
belonged to “Network Computing” with a publication level of 1230 published papers
during this period of time. The biggest bubbles represent the most published aspects
at that time such as “Optical Networking”, “Virtual Private Network”, “Ad Hoc
Network”, “Quality of Service”, “Internet Security” and “Wireless Internet”. Smaller
bubbles represent less published aspects such as “Peer-to-Peer Network”, “Network

Convergence”, “Cognitive Network” and “Fault Tolerant Network”. More
surprisingly, the idea of “Next Generation Network” existed already with 250
published papers by end of year 1999.
Fig. 2. FI Research Domain Landscape 1990-1999
On the vertical dimension from wired to wireless Internet, the island in the bottom
area is constituted of “Optical Networking” while the island in the top area is based
on “Wireless Internet”. The islands on the right and left hand spaces as well as the
islands in the bottom and top spaces are supposed to generate a certain gravity
attracting other concepts through the other dimensions of internet routing, evolution
approach and autonomic & convergent network.
All low brightness small bubbles represents emerging aspects with very few
published papers such as “IP Multimedia System” counting 3 published papers and
“Internet of Things” with 8 published papers. The only concept that was not emerging
by year 1999 is represented by “Content Centric Networking”, which scored 0
published papers, has a very low brightness level in the figure.

in Figure 3 where the concept of “Network Computer” forms a bigger island due to an
increased publication level of 19200 published papers. The opposite island,
constituted by the concept of “Network Computing”, is in this period only 4 times
smaller due to a double amount of 5860 published papers in the same period.
Interestingly, “Wireless Internet” has considerable grew up to 12400 published

papers as well as “Optical Networking”, hence, both generate much more gravity and
attraction. For example the concept of “Wireless Sensor Network” has turned from an
emerging bubble into a real one with 1140 published papers. However, in this same
period of time, “Internet of Things” has only scored 117 published papers.
Other concepts have turned from emerging concepts to confirmed ones, such as
“Semantic Web Services”, “Quality of Experience”, “Internet of Things”, Cognitive
Network”, “Autonomic Network”, “Self Adaptive Network” and “Resilient Network”
as well as “Cloud Computing” that respectively scored in Google Scholar 2390, 672,
117, 377, 151, 70, 179 and 127 published papers.
Similarly to the previous period of time, “Content Centric Networking” has still a
very low brightness level in the figure because it scored only 1 published paper,

in Figure 4 where the concept of “Network Computer” starts to decrease with a
publication level of 18100 published papers. The opposite island, “Network
Computing”, increases with 6960 published papers. This might be highlighting the
current transition from network computer towards network computing. The “Content-
centric Networking” concept emerges in this period with 81 published papers.
The most impressive progression comes from the concept of “Cloud Computing”
that exponentially moves from 127 published paper in the previous period
(2000-2005) up to 10200 in this period (almost factor 100). The concept of “Internet
of Things” moves in the same way but with a less exponential (factor 20) progression
from 117 to 2400. There are other concepts that make a good progression in this
period, such as “Autonomic Network”, “Wireless Sensor Network”, “Cognitive
Network” and “Quality of Experience”. Finally, the concept of “Next Generation
Network” makes also a significant progression in moving from 1650 to 4030
published paper in the period 2006-2011.
As for the concepts having a stable publication level over the two period
2000-2005 and 2006-2011, “Optical Networking” and “Internet Security” have a very
small increase of respectively 2% and 13%. The concept of “Quality of Services” is
also on the way to reach a stable plateau kind of situation with 20% progression
compare to the previous progression (5 times) from 1990-1999 to 2000-2005
publication levels.
3 Evolution of Interest in FI Networking Research Areas

A ranking from 1 to 24 was computed for each period based on the amount of
published papers. The evolution of interest in FI research areas reflected by the
publication weight of the respective FI concepts appears in Table 2, which highlights
the concept or research area ‘Pervasive Computing’ as the most popular for the time
period 2006-2011 and ‘Content-Centric Networking’ as the less popular.
Table 2. Ranking of Concepts along the different time-periods
Rank
Concepts
2006-2011 2000-2005 1990-1999
Content-centric Networking 24 24 24
Self-adaptive network 23 23 18
Resilient Network 22 19 17
Fault tolerant network 21 18 6
Autonomic Network 20 20 19
Cognitive Network 19 17 9
Network Convergence 18 14 5
Quality of Experience 17 15 15
Internet of Things 16 22 22
Optical Networking 15 10 10
IP Multimedia Subsystem 14 16 23
Next Generation Network 13 12 11
Peer-to-Peer Network 12 9 14
Quality of Services 11 11 3
Internet Security 10 7 1
Wireless Sensor Network 9 13 20
Semantic Web Services 8 8 21
Virtual Private Network 7 6 4
Cloud Computing 6 21 12
Wireless Internet 5 3 7
Ad hoc Network 4 5 8
Grid Computing 3 4 16
Ubiquitous Computing 2 1 2
Pervasive Computing 1 2 13
It shows as well that several FI research areas (concepts) while they were part of
the most popular in the time period 1990-1999, became the less popular in the time
period 2006-2011, such as ‘Fault Tolerant’ (from rank 6 to rank 21), ‘Network
Convergence’ (from rank 5 to rank 18), ‘Cognitive Network’ (from rank 9 to rank
19), ‘Quality of Services’ (from rank 3 to rank 11) and finally ‘Internet Security’
(from rank 1 to rank 10).
Others remain in the most popular, such as ‘Ubiquitous Computing’ (from rank 2
to rank 2), ‘Ad hoc Network’ (from rank 8 to rank 4), and ‘Wireless Internet’ (from
rank 7 to rank 5). Finally, FI research areas that were the less popular in the time
period 1990-1999, became the most popular in the time period 2006-2011, such as
‘Grid Computing’ (from rank 16 to rank 3), ‘Pervasive Computing’ (from rank 13 to
rank 1), ‘Cloud Computing’ (from rank 12 to rank 6), ‘Semantic Web Services’ (from
rank 21 to rank 8) and ‘Wireless Sensor Network’ (from rank 20 to rank 9).
Table 3. Ranking Evolution of FI Concepts in Two Time-Periods
Ranking Evolution
Concepts 2011-2006 1990-1999
Content-centric Networking 0 0
Self-adaptive network 0 -5
Resilient Network -3 -2
Fault tolerant network -3 -12
Autonomic Network 0 -1
Cognitive Network -2 -8
Network Convergence -4 -9
Quality of Experience -2 0
Internet of Things 6 0
Optical Networking -5 0
IP Multimedia Subsystem 2 7
Next Generation Network -1 -1
Peer-to-Peer Network -3 5
Quality of Services 0 -8
Internet Security -3 -6
Wireless Sensor Network 4 7
Semantic Web Services 0 13
Virtual Private Network -1 -2
Cloud Computing 15 -9
Wireless Internet -2 4
Ad hoc Network 1 3
Grid Computing 1 12
Ubiquitous Computing -1 1
Pervasive Computing 1 11
Interestingly, several FI networking research areas have a non homogeneous

ranking in the middle time period, such as ‘Cloud Computing’ with a transition from
rank 12 (1990-1999) to 21 (2000-2005) and finally 6 (2006-2011). Most of the other
research areas display a progression or regression. The most constant is the research
area ‘Ubiquitous Computing’ with the ranking 2-1-2 and ‘Next Generation Network’
with ranking 11-12-13 as well as ‘Autonomic Network’ with ranking 19-20-20.
Another way of looking at the ranking of FI networking research areas consists in
considering the number of lost/gained positions between 1 and 24 within the time
periods of 2006-2011 and 1990-1999 compared to the middle one 2000-2005. The
following table highlights the FI networking research areas with the highest gain, such
as ‘Cloud computing’ with a considerable gain of 15 position and ‘Internet of Things’
with a less impressive gain of 6 position (see Table 3). During the previous period
(1990-1999), the winning three were Semantic Web Services, Grid Computing and
Pervasive Computing.
4 Remarks and Conclusions

After digging into all these figures, one might be thinking that Cloud Computing and
Internet of Things research publication streams will continue to exponentially grow
due to the current interest in developing innovative services based on the ‘open data’
strategy developed by cities that deploy more and more sensors in their urban areas.
Surprisingly, Content Centric Networking appears to be quite flat in terms of
publication stream while current networks mainly move content objects through not
optimum host-to-host conversations. Is it due to a lack of research projects in the
CCN area? Or is it simply due to the anticipated deployment difficulties?
The lack of domain landscape on FI networking research appears to be a potential
topic of interest for researchers for the elaboration of the FI roadmap and related
networking research challenges for the next 10 years. It would help to reach a broader
understanding of the location and articulation of the various networking concepts.
Exploring the research domain landscape of FI and identifying related networking
concepts in digging into a large amount of published papers was a demanding but
useful task. The design of the FI landscape with concept bubbles inhabiting the
various territories reveals to be more fascinating. During this work, a territory of
“computing” emerged as a necessary bubble linking more recent concepts.
Developing a landscape on a wider FI domain could be useful for the research
community in order to identify and locate FI related concepts within dimensions
showing possible directions of progress. However, it would logically require the
participation and contribution of the whole FI research community. We hope that this
first tentative and issued FI landscape of networking concepts will motivate enough
other researchers for contributing to its future development.
Acknowledgements. This work was performed within the FIREBALL ICT project
and partly funded by the European Commission. The authors wish to acknowledge the
European Commission and project partners for their support. We also wish to
acknowledge our gratitude and appreciation to INRIA scientific leaders of project
teams involved in FI related projects, namely: Denis Caromel and Franca Perrina
(TEFIS), Walid Dabbous (PlanetLab and OneLab), Eric Fleury (SensLAB), and
David Margery (BonFIRE), for their active contribution.
References
1. Pirolli, P., Preece, J., Shneiderman, B.: Cyberinfrastructure for Social Action on National
Priorities. IEEE Computer Journal, special Issue on Technology-Mediated Social
Participation (November 2010)
2. Tselentis, G., Galis, A., Gavras, A., Krco, S., Lotz, V., Simperl, E., Stiller, B., Zahariadis,
T.: Towards the Future Internet - Emerging Trends from European Research (2010),
http://www.booksonline.iospress.nl/Content/View.aspx?piid=16
465 (retrieved)
3. Burin des Rosiers, C., Chelius, G., Fleury, E., Fraboulet, A., Gallais, A., Mitton, N., Noël,
T.: SensLAB Very Large Scale Open Wireless Sensor Network Testbed. In: Korakis, T.,
Li, H., Tran-Gia, P., Park, H.-S. (eds.) TridentCom 2011. LNICST, vol. 90, pp. 239–254.
4. Kaafar, M., Mathy, A., Barakat, L., Salamatian, C., Turletti, K., Dabbous, T.: Securing
Internet Coordinate Embedding Systems. In: Proceedings of ACM SIGCOMM 2007,
Kyoto, Japan (August 2007)
5. Schaffers, H., Santoro, R., Sällström, A., Pallot, M., Trousse, B., Hernandez-Munoz, J.M.:
Integrating Living Labs with Future Internet Experimental Platforms for Co-creating
Services within Smart Cities. In: Proceedings of the 17th International Conference on
Concurrent Enterprising, ICE 2011 Innovating Products and Services for Collaborative
Networks, Aachen, Germany (June 2011)
6. Leguay, J., Sallstrom, A., Pickering, B., Boniface, M., Gracia, A., Perrina, F., Giammatteo,
G., Roberto, J., Campowsky, K.: Testbed Facilities for Multi-faceted Testing throughout
the Service Development Lifecycle. In: ServiceWave 2011, Poznan (October 2011)
7. Friggeri, A., Chelius, G., Fleury, E., Fraboulet, A., Mentré, F., Lucet, J.-C.: Reconstructing
Social Interactions Using an unreliable Wireless Sensor Network. Computer
Communications 34(5), 609–618 (2011a)
8. Friggeri, A., Chelius, G., Fleury, E.: Egomunities, Exploring Socially Cohesive Person-
based Communities, N° RR-7535 (2011) [inria-00565336 - version 2] (2011b)
9. Pallot, M., Trousse, B., Senach, B., Scapin, D.: Living Lab Research Landscape: From
User Centred Design and User Experience towards User Cocreation. In: Proceedings of the
Living Lab Summer School, Cité des Sciences, Paris (August 2010)
Author Index
Angelou, Marianna 225 Gimenez, Diego 190

Ansaldi Oliva, Gustavo 81 Gimenez, Roberto 190
Anthopoulos, Leonidas G. 178 Gonçalves, Maurı́cio Amaral 142
Antoniou, Athanasios 117 Grammatikou, Mary 213
Autili, Marco 81 Grimm, Paul 130
Axenopoulos, Apostolos 130 Grosso, Paola 213
Grzech, Adam 154
Belter, Bartosz 213
Ben Hamida, Amira 81 Hjorth, Isis 42
Bertolino, Antonia 81
Bezzi, Michele 105 Issarny, Valérie 81
Boniface, Michael 42
Bourse, Didier 18 Jennings, Edel 30
Brzostowski, Krzysztof 154 Juszczyszyn, Krzysztof 154
Camargos, Lásaro Jonas 142 Kalatzis, Nikos 30

Campanella, Mauro 213 Kalogiros, Costas 6
Camurri, Antonio 130 Kaluvuri, Samuel Paul 105
Carabaño, Mario 190 Keim, Daniel 93
Chatzigiannakis, Ioannis 117 Kofuji, Sergio Takeo 142
Cobo, Cristobal 42 Kohlhammer, Jörn 93
Croce, Vincenzo 130 Kon, Fabio 81
Kosmides, Pavlos 30
Daras, Petros 130
Kostopoulos, Alexandros 6
Darmois, Emmanuel 18
Davey, James 93 Lazzaro, Marilena 130
de Andrade, Isabelle Cecı́lia 142 Levä, Tapio 6
De Angelis, Guglielmo 81 Liampotis, Nicolas 30
Demeester, Piet 55, 68 Lorenzo Toledo, Rubén M. 225
de Miguel, Ignacio 225 Lorré, Jean-Pierre 81
Denazis, Spyros 201 Lotz, Volkmar 105
Dias, Alex 142 Lymberopoulos, Leonidas 213
Di Cerbo, Francesco 105
Doolin, Kevin 30 Maglaris, Vasilis 213
Dos Santos, Carlos Eduardo Moreira 81 Malassiotis, Sotiris 130
Drapala, Jaroslaw 154 Mansmann, Florian 93
Durán Barroso, Ramón J. 225 Martin, Emilio 190
Martinez-Julia, Pedro 55
Etzold, Jonas 130
Massari, Alberto 130
Fekete, Attila 213 McKitterick, David 30
Ferreira, Caio César 142 Meyer, Eric T. 42
Fuentes, Diego 190 Morreale, Vito 55
Mylonas, Georgios 117
Gavazzi, Roberto 190
Georgantas, Nikolaos 81 Oostveen, Anne-Marie 42
252 Author Index
Pallot, Marc 237 Tafur Monroy, Idelfonso 225

Papadimitriou, Dimitri 18, 55, 68 Taylor, Nick 30
Papadopoulou, Elizabeth 30 Theodoridis, Evangelos 117
Papafili, Ioanna 6, 55 Theodoro, Luiz Cláudio 142
Papanikolaou, Nick 1 Tomkos, Ioannis 225
Pasquini, Rafael 142 Torelli, Francesco 55
Pereira, Fabı́ola Souza Fernandes 142 Trabelsi, Slim 105
Pereira, João Henrique de Souza 142 Tranoris, Christos 201
Pertejo, Judith 190 Trossen, Dirk 6
Pickering, Brian 42 Trousse, Brigitte 237
Potts, Martin 213 Tsekeridou, Sofia 190
Tymecki, Andrzej 225
Roddy, Mark 30
Tzovaras, Dimitrios 130
Rosa, Pedro Frosi 142
Roussaki, Ioanna 30
Vakali, Athena 166, 178
Sabetta, Antonino 105 Venâncio Neto, Augusto José 142
Sales, Bernard 18, 55, 68 Virgos, Sofia 190
Salvadori, Elio 225
Santos, Eduardo de Souza 142
Wainwright, Nick 1
Schroeder, Ralph 42
Senach, Bernard 237
Silva, Flávio de Oliveira 142 Ye, Yabin 225
Siracusa, Domenico 225
Srivastava, Lara 166 Zahariadis, Theodore 55, 68
Steiner, Thomas 130 Zarras, Apostolos 81
Światek,
Pawel 154 Zhang, Nan 6

2012 Book TheFutureInternet

Uploaded by

Copyright:

Available Formats

2012 Book TheFutureInternet

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

2012 Book TheFutureInternet

Uploaded by

Copyright:

Available Formats

Lecture Notes in Computer Science 7281

Commenced Publication in 1973

The Future Internet

Future Internet Assembly 2012:

Federico Álvarez Stamatis Karnourskos Hans Schaffers

Acknowledgement and Disclaimer

ISSN 0302-9743 e-ISSN 1611-3349

Library of Congress Control Number: 2012937591

CR Subject Classification (1998): C.2, D.4.4, D.2, H.3.5, H.4, K.6.5

LNCS Sublibrary: SL 3 – Information Systems and Applications, incl. Internet/Web

There were 40 submissions. Each submission was peer-reviewed by experts

March 2012 Federico Álvarez

in the scale of magnitude. Some other limits are determined by non-common

systems, addressing scalability and heterogeneity issues in order to leverage

puting resources, will have considerable impact as it is opening up new possibili-

2 Papers in the Section Applications

retrieval of diverse media types simultaneously e.g. 3D objects, images, audio

Invited Introduction Paper

A Systematic Approach for Closing the Research to Standardization

SOCIETIES: Where Pervasive Meets Social . . . . . . . . . . . . . . . . . . . . . . . . . 30

Cross-Disciplinary Lessons for the Future Internet . . . . . . . . . . . . . . . . . . . . 42

Design Principles for the Future Internet Architecture . . . . . . . . . . . . . . . . 55

From Internet Architecture Research to Standards . . . . . . . . . . . . . . . . . . . 68

An Integrated Development and Runtime Environment for the Future

Visual Analytics: Towards Intelligent Interactive Internet and Security

Using Future Internet Infrastructure and Smartphones for Mobility

I-SEARCH: A Uniﬁed Framework for Multimodal Search and

Semantically Enriched Services to Understand the Need of Entities . . . . . 142

Supporting Content, Context and User Awareness in Future Internet

Urban Planning and Smart Cities: Interrelations and Reciprocities . . . . . 178

The Safety Transformation in the Future Internet Domain . . . . . . . . . . . . 190

NOVI Tools and Algorithms for Federating Virtualized

Next Generation Flexible and Cognitive Heterogeneous Optical

A Tentative Design of a Future Internet Networking Domain

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Nick Wainwright and Nick Papanikolaou

Cloud and Security Lab, HP Labs

Abstract. We describe the key ﬁndings of the Future Internet Assem-

Research Priorities. In presenting the research needs in this roadmap we have

1. Going Beyond Converged Infrastructure: the Internet infrastructure

4. Networked Interaction: people interacting with each other, with infor-

Going Beyond Converged Infrastructure. Internet infrastructures — net-

– Polymorphic networks — combining diﬀerent networks to meet the for ca-

The Rise of Networked Data. Networked data on a massive scale is the

Achieving Real Internet Security. A decade away we will be conducting

unobtrusive, so that they do not hinder openness and availability of network

Networked Interaction. Through the Internet we interact with each other,

Augmented Worlds. The vision of networked services, systems, and devices

Internet-style Innovation. It is clear that the Internet has been an incredible

Conclusions. We have highlighted the themes above because they need to be

Alexandros Kostopoulos1, Ioanna Papafili1, Costas Kalogiros1, Tapio Levä2,

Keywords: information-centric networking, content delivery, future internet

2 A Methodology for Tussle Analysis

2. Identify tussles among identified stakeholders.

Several techniques or methods can be used to perform each of the aforementioned

Fig. 2. Notation of VNC methodology

3 Overview of ICN Architectures

Fig. 3. A Publish/Subscribe architecture for ICN [13]

At the application level, the pub/sub components implement applications based on

3.2 Network of Information