Short Paper: IoT: Challenges, Projects, Architectures

Vangelis Gazis, Manuel Goertz, Marco Huber, Alessandro Leonardi,

Kostas Mathioudakis, Alexander Wiesmaier, Florian Zeiger
AGT Group (R&D) GmbH, Hilpertstrasse 35
64295, Darmstadt, Germany
{vgazis,mgoertz,mhuber,aleonardi,kmathioudakis,awiesmaier}@agtinternational.com, fzeiger@ieee.org

Abstract—Internet of Things (IoT) is a socio-technical phenom- A. Application Domains

ena with the power to disrupt our society such as the Internet
before. IoT promises the (inter-) connection of myriad of things
We believe that IoT is application-driven, hence applica-
proving services to humans and machines. It is expected that tion requirements will boost the innovation and development
by 2020 tens of billions of things will be deployed worldwide. It of IoT. Major domains identified are: Energy, Smart City,
became evident that the traditional centralized computing and Transportation, Smart Home, Environment, Supply Chain, and
analytic approach does not provide a sustainable model this new Health Care. Nevertheless, this is only an indicative collection
type of data. A new kind of architecture is needed as a scalable
and trusted platform underpinning the expansion of IoT. The
of the most well-known uses cases, based on our research.
data gathered by the things will be often noisy, unstructured B. Challenges
and real-time requiring a decentralized structure storing and
analysing the vast amount of data. Several research and technology challenges need to ad-
In this paper, we provide an overview of the current IoT dressed towards the implementation of IoT applications as well
challenges, will give a summary of funded IoT projects in Europe, as the potential realization of horizontal IoT platforms. The
USA, and China. Additionally, it will provide detailed insights
most important challenges, based on our findings, are listed
into three IoT architectures stemming from such projects.
below.
I. I NTRODUCTION 1) Technological Interoperability: Interoperability is sig-
IoT will become the technological innovation driving ap- nificantly more challenging for the IoT as it is not (only)
plications that have the power to change the markets across about connecting people with people, but about a seamless
different domains. Thousands of applications can be identified interaction between devices and people with devices. These
in each domain and new ones appear everyday, requiring devices can differ regarding their technological capabilities.
a strong interconnection among things [1]. Interconnection 2) Semantic Interoperability: For full interoperability, it is
is not only a mere technological issue but it concerns also necessary that the devices interpret the shared information
aspects such as privacy, standardization, legal issues, etc. correctly and act accordingly, which is covered by the semantic
This inevitably brings new challenges driving research and aspect of interoperability usually referred to as Information
innovation in industry and academia over the last decade [2]. Model. Hence, improvements have to be made regarding
We believe that the core technological challenges, such distributed ontologies, semantic web, or semantic device dis-
as interconnection among heterogenous devices, very low covery.
computational and energy demand have to be overcome to 3) Security and Privacy: Data integrity, unique identifica-
pave the road for the adoption of IoT. Projects in industry tion, and encryption are considered core challenges for IoT, as
and academia around the world strive to solve parts of these much of the data being acquired and communicated contain
challenges. A fundamental importance will be the development personal information. Additionally, data ownership, legal and
of an open, scalable and trusted architecture. liability issues have to be addressed accordingly. Finally,
The rest of the paper is organized as follows: Section energy efficient encryption and data protection technologies
II identifies the technological challenges of IoT and the have to be considered.
most promising application domains. Section III provides an 4) Smart Things: Ultra low power circuits and devices
overview of the most important public funded projects in capable of tolerating harsh environments have to be developed.
Europe, in the United States, and China. Three promising Moreover, parallel processing in low power multi-processor
IoT architectures developed in public funded projects will be systems, adaptation, autonomous behavior while guaranteeing
covered in detail in Section IV. Finally, Section V closes the trust, privacy and security, as well as battery, energy harvest-
paper with conclusion and outlook. ing and storage technologies are among the core challenges
regarding the devices in the IoT.
II. A PPLICATIONS AND C HALLENGES 5) Resilience and Reliability: In industrial environments or
Major objectives for IoT are the creation of smart envi- in emergency use cases temporary outages cannot be accepted.
ronments/spaces and self-aware things. In this Section we Hence, resilience and reliability issues in IoT need to be
consolidate application domains and challenges will drive the investigated from an overall systems view and in addition
evolution of IoT systems. comprise aspects like availability, robustness and flexibility of
 Interoperability

Interoperability
Technological

Smart Thing
Architecture

Reliability
Resilience
Semantic

Security
Privacy
Acronym
IoT-A [4] w l l l
Butler [5] w w l w
IoT.est [6] w l w
RELYonIT [7] w l w l
IoT6 [8] l w w
IoT Lab [9] w l w
uTRUSTit [10] w l l
BETaaS [11] l l l l l
iCORE [12] l l w
OPENIoT [13] l l l w w
CityPulse [14] l l l
TABLE I
C APABILITIES ADDRESSED IN EU I OT PROJECTS .
Fig. 1. Clusters of NSF CPS projects.

the communication and hardware to changing environmental

conditions, avoidance of single points of failure, or the robust-
ness of data processing to uncertain information.

III. P ROJECTS
The IoT paradigm has been picked up by governments
around the world and funding schemes have been imple-
mented. Next we will list our findings analysing the funding
schemes of the EU, USA, and China.
Fig. 2. Clusters of China IoT projects.
A. EU funded projects
The European Union has set up the European Research C. Chinese IoT topics
Cluster on the Internet of Things funding 33 projects in IoT is one of the major technology topics and part of the
total [3]. Table I lists the most recognized projects and shows 12th 5-year-plan of China. The funding is clustered among a
to what extend (full (l), partly (w), none ( )) they address the layered schema: Perception Layer, Network Layer, Transport
challenges described in Section II. Layer, and Application Areas. A strong focus lies on network
A focus was given to the development of open, scalable, and technology supporting IoT application. A summary of identi-
reliable IoT architectures. Section IV of this paper will give a fied IoT sectors in China is depicted in Figure 2.
detailed view on the findings of the projects IoT-A, BETaaS,
and OPENIoT regarding IoT architectures. IV. A RCHITECTURES
A scaleable and reliable architecture will form the backbone
B. NSF funded projects
of the future development of IoT. The architecture must cope
The President’s Council of Advisors on Science and Tech- with the new requirements of IoT and reflect the challenges
nology published in December 2010 the report "Designing a listed in Section II. This Section covers the architecture ref-
Digital Future: Federally Funded Research and Development erence model IoT-A, as well as two full-fledged architectures
in Networking and Information Technology". It provides the developed in the projects BETaaS and OPENIoT.
base for investment into research on Cyber-Physical-Systems
managed by the The National Science Foundation (NSF). A. IoT-A
A primary focus lies on medical systems and projects in the The IoT Architecture Reference Model (ARM) [15] is not
area of software engineering. However, the clustering of all an IoT architecture per se, but a set of best practices, guide-
project has also identified the CPS sectors energy, smart cities lines, and a starting point to generate specific IoT architectures.
and buildings, manufacturing, intelligent traffic management, It provides an architectural reference model facilitating the
and disaster and hazards. interoperability of IoT systems. It also provides the tools, such
Most projects (cf. Figure 1) emphasize on fundamental re- as resolution, look up, and discovery of things, for the actual
search advancing CPS in its capability, adaptability, scalability, integration into the service layer.
resiliency, safety,security, and usability propelling CPS with its The ARM Process defines the steps to generate concrete IoT
tightly intertwined physical processes and networked comput- architectures from business goals, informing on IoT related
ing capabilities beyond todays simple embedded systems. issues in a methodology agnostic way. The covered topics
include the generation of requirements and their transforma-
Support for heterogeneous sensor network types and scal-
tion into an architecture using views and perspectives. ARM
ability to the proliferation of sensory devices is enabled by
provides an exhaustive list of so-called Unified Requirements
supporting a distributed deployment model for SM instances.
(UNIs) [16], that can be used to generate concrete require-
ments for a specific architecture. The UNIs are generalized
V. C ONCLUSIONS
requirements augmented with the views and perspectives of
the respective stakeholders. The current adoption of IoT will influence many application
The ARM process makes use of the IoT Reference Model areas of our society. In this paper the major challenges that
that introduces major IoT concepts such as devices, services, need to be tackled making IoT ubiquitous have been identified
and entities and defines their relations and attributes on an and laid out. The derived requirements are later mapped with
abstract level that is independent from specific use cases or the current research activities addressing specific parts of the
technologies. Following the established relations, it identifies challenges.
so-called Functional Groups (FGs) for interacting with in- Governments around the world, most notable US, Europe,
stances of the introduced concepts and introduces communi- and China have put IoT as one the top priorities on their re-
cation functionalities suitable for heterogenous IoT settings. search agenda. An overview on current public funded projects
Additional features introduced include trust, security, process in these three geographic regions has been given.
management, service organization, and more. A saleable, reliable and trusted architecture is essential to
B. BETaaS allow IoT to become a world-wide business critical infrastruc-
ture. Three major projects developing IoT architectures have
Building the Environment for the Things as a Service being identified and their results have been described.
(BETaaS) [17] defines besides the overall functionality and
architecture, an actual implementation of the platform is part R EFERENCES
of the deliverables. BETaaS is a running project and some
deliverables are not yet available in their final version or are [1] Future Internet: The Internet of Things Architecture, Possible
Applications and Key Challenges, 2012. [Online]. Available: http:
even still missing completely. //ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=6424332
BETaaS consists of a network of gateways (“local cloud of [2] I. G. Smith, O. Vermesan, P. Friess, and A. Furness, The
gateways”) that seamlessly integrate existing heterogeneous Internet of Things 2012 New Horizons, I. G. Smith, Ed.,
2012. [Online]. Available: www.internet-of-things-research.eu/pdf/
M2M systems. To abstract from the heterogeneity of the IERC_Cluster_Book_2012_WEB.pdf
physical layer, BETaaS defines and builds upon a baseline [3] “Internet of things, pan european research and innovation vision,”
reference architecture called Things-as-a-Service (TaaS). The 2011. [Online]. Available: http://www.internet-of-things-research.eu/
TaaS Reference Model is the foundation of the BETaaS pdf/IERC_IoT-Pan%20European%20Research%20and%20Innovation%
20Vision_2011_web.pdf
infrastructure on which it is built. It provides architectural [4] M. Rossi (ed.), “IoT-A WP3 – Protocol Suite,” http://www.iot-a.eu/
models for domains, information, communication, security, public/public-documents/documents-1/1/1/d3.3/at_download/file, April
and functions. TaaS builds upon the IoT-A models (Sec. IV-A), 2012, [04 Feb. 2014].
[5] “BUTLER project description.” [Online]. Available: http://www.
tailoring and extending them to its specific needs. iot-butler.eu/
[6] “IoT.est project description,” 2013. [Online]. Available: http://ict-iotest.
C. OPENIoT eu/iotest/
The OPENIoT research project has defined an architecture [7] “RELYonIT project description,” 2013. [Online]. Available: http:
//www.relyonit.eu/
utilizing a Sensor Middleware (SM) and a Semantic Direc- [8] “IoT6 project description,” 2012. [Online]. Available: http://www.iot6.
tory Service (SDS) [18]. To achieve alignment, architecture eu/
development and specification was based on the Architecture [9] “IOT LAb pproject description,” 2013. [Online]. Available: http:
//www.iotlab.eu/index.php
Reference Model (ARM) of IoT-A. [10] “UTRISTIT project description,” 2013. [Online]. Available: http:
The Sensor Middleware (SM) undertakes the collection, //www.utrustit.eu/
filtering and aggregation of data streams associated to physical [11] “BETaaS project description,” 2013. [Online]. Available: http://www.
betaas.eu/
and virtual objects. The Cloud Computing Infrastructure (CCI) [12] “iCore project description,” 2013. [Online]. Available: http://www.
supports the storage of data along with their associated meta- iot-icore.eu/
data information in a scalable and elastic manner. The Seman- [13] “OPENIoT project description,” 2013. [Online]. Available: http:
//www.openiot.eu/
tic Directory Service (SDS) supports registration management [14] “City Pulse project description,” 2013. [Online]. Available: http:
and semantic annotation for sensors and services. The Global //www.ict-citypulse.eu/
Scheduler (GS) is tasked with handling requests for on-demand [15] A. Bassi, M. Bauer, M. Fiedler, T. Kramp, R. Kranenburg, S. Lange,
service deployment and the associated provisioning of access and S. Meissner, Eds., Enabling Things to Talk. Springer Heidelberg
New York Dordrecht London, 2013.
to data sets and services that may be required. The Request [16] I.-A. Consortium, “Iot-a unified requirements,” http://www.iot-a.eu/
Definition (RD) element enables the dynamic specification of public/requirements/, [31 Jan. 2014].
service requests and the Request Presentation (RP) element [17] BETaaS Consortium, “Building the environment for the things as a
service,” http://www.betaas.eu/, 2012, [11. Feb. 2014].
is tasked with the visualization of the results produced by an [18] “OPENIoT project description,” 2013. [Online]. Available: http:
executing service. //www.openiot.eu/