ZORZI LAYOUT 12/9/10 11:15 AM Page 44

THE INTERNET OF THINGS

FROM TODAY’S INTRANET OF THINGS TO A

FUTURE INTERNET OF THINGS:
A WIRELESS- AND MOBILITY-RELATED VIEW
MICHELE ZORZI, UNIVERSITY OF PADOVA
ALEXANDER GLUHAK, UNIVERSITY OF SURREY
SEBASTIAN LANGE, VDI/VDE INNOVATION + TECHNIK GMBH
ALESSANDRO BASSI, HITACHI EUROPE, LTD.

Resolution infrastructure
ABSTRACT possible to interact with the environment around
us, and to receive information on its status that
Monitoring of
associations In this article, we present the current status was previously not available by simply looking at
between entities of the Internet of Things, and discuss how the things. Moreover, it will be possible not only to
and IoT resources
current situation of many “Intranets” of Things actively interconnect things in the physical world,
Monitoring

In should evolve into a much more integrated and but also to enable them to exchange and make
heterogeneous system. We also summarize what use of their information (in the “digital” world).
in our opinion are the main wireless- and mobili- Given the current fragmentation of efforts in
ty-related technical challenges that lie ahead, this area, which prevents a synergistic integration
and outline some initial ideas on how such chal- process, we believe that there is a clear need to
lenges can be addressed in order to facilitate the develop a reference architectural model that will
Moving IoT resource IoT’s development and acceptance in the next allow interoperability between different systems.
few years. We also describe a case study on the The TCP/IP protocol suite, by now universally
IoT protocol architecture. used in the Internet, emerged from a pool of
many different protocols rather than “being cre-
M INTRODUCTION ated” following a clean-slate approach. Similarly,
instead of designing “the” IoT from scratch, we
The authors present The Internet of Things (IoT) arena as of today believe that a better approach will be to inte-
grate existing efforts and current technologies
resembles the “Wild West” of a couple of cen-
the current status of turies ago. It is a vast, mostly unexplored territo- into a unified, ubiquitously applicable architec-
ry, without clear borders, where all current ture, to design open and flexible interfaces and,
the Internet of technologies can play a role, and where ad hoc where missing, to develop the necessary bridges
solutions are often the norm. Governance is very to connect different technologies from the hard-
Things, and discuss limited and contradictory, and attacks, both from ware to the service layer. We firmly believe that
consumer associations scared of a “big brother only such a foundational work will jump start the
how the current like” control and from business executives scared Internet of Things era, whose impact on our
situation of many of new and radically different business models,
can seriously hamper its development.
lives can be expected to be even more dramatic
than what the Internet has brought about.
“Intranets” of Things But, as the Wild West, the mirage of the With respect to the technological roadblocks
seemingly unlimited capabilities offered by the (the focus of this article), we see an immediate
should evolve into “web of things” is driving more and more need for action in three different areas:
research and development solutions in this area. 1. An architectural reference model for the inter-
a much more IoT is seen as a pillar of the Future Internet [1], operability of IoT systems, outlining principles
and recent advances in battery life, miniaturiza- and guidelines for the design of its protocols,
integrated and tion, energy harvesting, transmission protocols, interfaces, and algorithms
heterogeneous and hardware costs are bringing its vision closer
and closer to reality.
2. Mechanisms for the efficient integration of
this architecture into the service layer of the
system. The IoT vision [2, 3] of pervasively connecting Future Internet networking infrastructure
smart things will provide a unique chance to 3. A Novel resolution infrastructure, allowing
enable a rich set of evolutionary as well as revolu- scalable lookup and discovery of IoT
tionary applications and services (see [4] for some resources, entities of the real world, and their
interesting examples). For the first time it will be associations

44 1536-1284/10/$25.00 © 2010 IEEE IEEE Wireless Communications • December 2010

ZORZI LAYOUT 12/9/10 11:15 AM Page 45

INTRAnets of things INTERnet of things

Home
and health
Sensor devices Sensor
networks networks
Blue-
tooth Actuation
IP networks Home
IP6lo
Actuation wpan and health
networks devices
Propretory IoT’s
arch.
NFC
ISO
Enter- 14443 Enter-
tainment tainment
networks NFC networks

x Protocols and technologies (e.g.) driving the intranets of things Interconnectivity by IoT’s architecture

Figure 1. From the current “Intranets of Things” to the “Internet of Things.”

This architectural framework will make it retail and logistics, it is foreseen that the appli-
possible to overcome the current fragmentation cation of RFID tags to any object around us will
and limitation of solutions, where many open up the possibility to develop a huge num-
“Intranets” of Things exist, towards a true ber of disruptive services. Although some privacy
“Internet” of Things, where all devices will be concerns have been raised, RFIDs have become
part of a globally integrated system. (A pictorial part of our everyday life. Despite the recent
view of this concept is provided in Fig. 1.) In this improvements in RFID technology (e.g., in
article, we will discuss some of the main techni- terms of miniaturization), further developments
cal challenges related to wireless communica- need to be realized, especially in the areas of
tions and mobility in an IoT context, and outline energy harvesting and batteries, integration into
some initial ideas towards possible solutions. materials, and cost. Furthermore, integration of
RFIDs in other devices such as mobile phones
RELATED WORK and sensors, which is seen as a major step
towards the development of IoT, is not yet
The term “Internet of Things” was coined more widespread.
than 10 years ago by the Auto-Id Labs in the US With respect to numbering and identification
(http://www.autoidlabs.org/), where in parallel issues, organizations such as EPC Global
the concepts of “ambient intelligence” and (www.epcglobalinc.org), the IP for Smart Objects
“ubiquitous computing” were also being devel- Alliance (www.ipso-alliance.org) and the Ubiqui-
oped. Since then, there have been some consid- tous ID Center (www.uidcenter.org) currently
erable developments, in both academia and aim at promoting standards, in order to make
industry, in the US as well as in Europe and RFID interoperable and ubiquitously applicable.
Asia. Such developments have primarily been While these efforts are certainly important and
dedicated to applying RFID technology to the worthwhile, there are some concerns about their
logistics value chain [5]. Beyond this, sensor net- impact on how the field will evolve and their vul-
works have been applied in numerous industrial nerability to possibly unfair practices [8].
environments for process monitoring [6]. The Sensor networks, that are considered another
first trials to establish IoT-like applications for pillar of the IoT, experienced a similar develop-
end users in public have been set up in so called ment in the recent past. A lot of research has
“future stores” in Germany, Switzerland and been done in the last couple of decades, not only
Japan. Despite these early efforts, some of the in the area of networking protocols, including
more visionary aspects of the IoT concept have MAC and routing, but also in micro and nano
only been studied in laboratory scale use cases, technologies, as well as on higher-layer issues,
and many more are still waiting to be developed such as middleware, security and applications
and invented. [9]. In particular, technological advances allow
The Internet of Things theme can be seen as to have extremely sensitive, extremely precise,
a concept originating from converging topics and extremely small sensors, combining different
that find their origin in a number of different sensing techniques to minimize reading errors
research streams [2]. Its cornerstone is undoubt- and calibration issues. Sensor networks and
edly the RFID technology. Radio Frequency RFID research are slowly growing together [10]
Identifiers (RFIDs) were first introduced to as sensor nodes are becoming smaller and more
overcome the limitations of the barcode technol- highly integrated and RFID tags are equipped
ogy and primarily focus on tagging objects by with more computing power and storage capaci-
attaching an individual identifier to them [7]. ty than required by just an identification code. In
While the original idea was to tag items for the future we may expect that smart functional

IEEE Wireless Communications • December 2010 45

ZORZI LAYOUT 12/9/10 11:15 AM Page 46

One of the key components such as tags and sensors will be

inherently integrated into the environment, so
task, especially with reference to wireless com-
munications devices with vastly different levels
requirements for the that their presence will be pervasive but invisi- of capabilities and performance, and will involve
ble. an inevitable performance degradation com-
IoT paradigm to be In the wireless communication area, it is pos- pared to a highly optimized vertical design.
sible to identify several parallel developments.
successful is its Efforts such as ZigBee, Bluetooth, Wi-Fi, NFC, CONNECTIVITY
all with their own specific characteristics and Another key area of investigation relates to how
ability to integrate application domains, have reached a significant to provide communications capabilities to the
many types of maturity and market size. However, the frag-
mentation of these protocols may again hamper
various devices involved, that in many cases will
be wireless. Issues such as communications ener-
devices, technologies, objects interoperability and can slow down the gy consumption, antenna design, interoperability
development of a unified reference model for of different technologies (e.g., via cognitive radio
and services. At the the IoT. capabilities), adaptive techniques for a dynamic
Today, while single applications are getting environment in the face of possibly heavily con-
device level, this more and more common, the necessity of a uni- strained resources, etc. will have to be addressed.
fied approach for the IoT is still a topic promot- It will be important to understand what needs to
includes very ed and developed by a fairly small group of be connected so as to provide the necessary
diverse features. enthusiasts from academia, industry and public
institutions. In order to promote IoT as a pub-
communications capabilities, while avoiding that
a system that is too connected becomes hard to
licly accepted and implemented paradigm, and to manage (e.g., due to excessive interference). In a
guarantee a sustainable development of the relat- more futuristic scenario, alternative means of
ed technologies and products, standardization communications may also be considered, includ-
will be crucial as it will guarantee that the ing for example communications through organic
unavoidable evolution towards smart communi- matter.
cating objects will lead to a unified and integrat-
ed system of heterogeneous devices (the SCALE
“Internet” of Things), as opposed to separate If all things need to be part of this system, anoth-
vertical “silos” solutions which, though possibly er major challenge is related to the sheer num-
highly optimized, cannot have a global scope (the bers involved. It is well known that in the
“intranets” of things). Some initial effort in this presence of many nodes, which cannot be a pri-
area has been made, notably within ETSI [11]; ori expected to be tightly coordinated, the per-
however, in order to achieve the necessary scale, formance of most communications schemes will
different areas (such as for example naming reso- suffer, assuming they can work at all. In a wire-
lution and the definition of common interfaces less context, such problems are further exacer-
for different device technologies) need to under- bated, e.g., due to the difficulty of having a
go a substantial standardization effort as well. coherent and stable view of the topology because
The key challenge in the years to come will of the channel- and mobility-related dynamics,
therefore involve both significant conceptual as well as the inherent unreliability of the medi-
innovation to solve the many open problems and um, which in dense multi-hop environments may
an effort to bring these concepts to real-life sys- become a serious bottleneck. Communications
tems and applications. protocols will therefore pose several challenges.
Management of the network becomes very diffi-
KEY TECHNICAL ISSUES cult in a large distributed environment, and solu-
tions to dominate the complexity need to be
Given the complexity of the IoT paradigm, the found. Despite the huge literature on these top-
list of challenges to be addressed is potentially ics, such extreme situations have not been ade-
very long. In the paragraphs below we will con- quately studied, and the behavior of the overall
centrate on the major technical roadblocks relat- system in terms of data handling performance,
ed to wireless communication and mobility. A stability, fairness, reliability, etc. will need to be
discussion on more general issues can be found revisited. The challenges related to other aspects
in [2]. (e.g., heterogeneity or addressing) will also be
exacerbated by the very large number of devices
HETEROGENEITY involved.
One of the key requirements for the IoT
paradigm to be successful is its ability to inte- NAMING, ADDRESSING AND IDENTIFICATION
grate many types of devices, technologies, and Identifying an object is one of the primary pillars
services. At the device level, this includes very of the IoT. A key problem is how to split Loca-
diverse features in terms of data communication tion and Identification of a device. Although
capabilities (e.g., data-rates and/or reliability), workarounds exist for IPv4, and some mecha-
computational and storage power, availability of nisms were introduced in IPv6 to support Inter-
energy, flexibility in handling different technolo- net mobility, the heterogeneity of existing
gies, mobility, etc. As to services, the system identification mechanisms and their co-existence
should ideally be open to supporting a huge vari- and efficient use across different systems make
ety of different applications whose characteris- this problem wider and more complex to solve
tics and requirements may be extremely diverse, for RFIDs. While the RFID world has been ini-
in terms of bandwidth, latency, reliability, etc. tially dominated by the use of EPC for identifi-
These heterogeneity traits of the overall system cation, uID [12] has recently gained popularity
make the design of a unifying framework and of as a more flexible alternative. Future identifica-
the communication protocols a very challenging tion schemes will have to embrace various enti-

46 IEEE Wireless Communications • December 2010

ZORZI LAYOUT 12/9/10 11:15 AM Page 47

ties of the real world such as places or living

beings, making the world-wide convergence to a
harvesting using energy conversion, as well as
extremely low-power circuitry and energy effi-
Issues such as
single scheme highly unlikely (not to mention cient architectures and protocol suites, will be medium access con-
the political considerations concerning gover- the key factors for the roll out of autonomous
nance). Suitable solutions therefore must be able wireless smart systems. trol, routing, error
to inter-relate heterogeneous schemes, consider-
ing also heterogeneous location/ID splitting TOWARDS A SOLUTION control, and Quality-
techniques, across different system boundaries.
There is little doubt that comprehensive of-Service will need
PRIVACY & SECURITY
One major social concern related to pervasive
research efforts in a variety of areas are
required, in order to come up with the neces-
to be explicitly
systems that have learning and reasoning capa- sary architectural foundations for the IoT. As a incorporated in the
bilities and can collect and store data about the focused contribution to the general discussion
environment is the way such data may be used on how the key technical IoT challenges are to design. Such issues are
(or abused). In addition, in applications where be addressed, in this section we concentrate on
the system is called to act on the physical reality, three wireless-related issues, and provide a case- made significantly
tampering with the system by a malicious intrud- study IoT architecture implementation as a con-
er may result in severe consequences in terms of crete example. more difficult than in
performance, operational disruption, theft, or traditional (wireless)
even safety hazards. In a wireless IoT context, COMMUNICATIONS AND NETWORKING
besides the obvious weakness of the radio chan-
IN DENSE ENVIRONMENTS networks by the
nel with regard to eavesdropping, the heavily
constrained nature of the devices and the limited Effective communications and networking in a heterogeneity and
bandwidth available make it very challenging to large and dense heterogeneous environment
provide effective security mechanisms via simple need protocols at the lower layers that support density of the
algorithms with limited room for message co-existence of diverse wireless interfaces, such
exchanges. We observe that these privacy and as intelligent (or cognitive) management of deployment.
security issues are not unique to this environ- interference and distributed management of
ment but need to be addressed in a variety of channel allocation/medium access. Communica-
systems, so that some concepts that already exist tion protocols exploiting the locality in order to
or are being developed in different contexts can deal with scale are a promising approach, as in
likely be reused. the IoT many interactions are expected to take
place locally between physically co-located
SELF-MANAGEMENT CAPABILITIES devices, and do not require support for global
In order to support the expected scale of the end-to-end interactions, as in the “traditional”
IoT, devices will need to self-manage without Internet. Where needed, end-to-end interactions
external intervention. Orchestration and manage- can be resolved by smart ways of aggregation,
ment mechanisms as well as information models e.g., by allowing nested aggregation of objects
will have to be defined taking into account this and associated IoT resources to deal with scale,
scale of deployment. The success of the Internet exploiting physical proximity and inter-relation-
lies in its minimalistic best-effort service ships between objects (for example, objects
approach. When trying to apply a similarly sim- placed in a container) for the use of addressing
ple approach to the IoT architecture, we have to or for discovery.
face the exponential growth in complexity that Issues such as medium access control, rout-
the connection of billions of heterogeneous ing, error control, and Quality-of-Service will
devices will bring, which will call for context need to be explicitly incorporated in the design.
awareness, self-organization, self-management, Such issues are made significantly more difficult
self-optimization, self-healing and self-protection than in traditional (wireless) networks by the
capabilities. In a wireless context, the increased heterogeneity and density of the deployment.
topology dynamics and likelihood of faults due to Given the potentially high level of dynamics in
channel fluctuations and possible device mobility, the system and the limited bandwidth, energy,
as well as the loss of signaling and control mes- storage and communications capabilities of most
sages, make these issues all the more challenging of the devices involved, the need to maintain
and call for schemes where robustness may be and process state should be kept to a minimum.
more important than efficiency. Also, robust schemes which provide the capabili-
ty to survive bad channel conditions or tempo-
ENERGY MANAGEMENT rary (or even prolonged) lack of connectivity will
Energy in all its phases (harvesting, conservation be of paramount importance in this context. In
and consumption) is a major issue, not only in our opinion, the following components will play
the IoT area, but more in general for the society a key role in developing a solution with the
at large. The development of novel solutions above features, towards a protocol suite able to
that maximize energy efficiency is paramount. In support IoT systems:
this respect, current technology is inadequate, •Random access and geographic routing:
and existing processing power and energy capaci- random access is inherently distributed, and suit-
ty are too low to cope with future needs. The able in an environment where knowledge of the
development of new and more efficient and topology is not always available or accurate; it
compact energy storage sources such as batter- can also be effectively coupled, via cross-layer
ies, fuel cells, and printed/polymer batteries, optimization, with geographic routing, which is
together with new energy generation devices stateless and can be made very efficient in dense
coupling energy transmission methods or energy topologies. In addition, we could also try to

IEEE Wireless Communications • December 2010 47

ZORZI LAYOUT 12/9/10 11:15 AM Page 48

One of the main fac- leverage on the presence of devices with differ-
ent capabilities, moving away from the “flat”
permits a wide range of different communication
implementations below, and a wide variety of
tors that contributed design concept typical of ad hoc networks, transport and application protocols above, allow-
towards a more efficient structure in which ing a highly diversified range of services and
to the success of the nodes take on roles according to their own capa- technologies to inter-operate. As the variety of
bilities and where more capable nodes may give technologies and services that belong to the IoT
current Internet archi- resource-poor devices access to advanced fea- is even wider, we believe that building an archi-
tures (e.g., computing or storage). This approach tecture that is centered on a similar concept will
tecture is the central- will open up several research challenges that do be essential for its success, allowing heteroge-
ity of the IP. As the not currently have a definitive solution.
•Network coding and random data combina-
neous technologies to thrive underneath and
applications and services to flourish above.
variety of technolo- tion: network coding has been shown to provide However, the complexity of the IoT and the
an effective means for efficient reliable data dis- heterogeneity of devices and technologies pose a
gies and services semination and to require little coordination much greater challenge than in the traditional
among nodes; random data combination is a Internet. The quest for the proper hourglass
that belong to the lightweight, yet effective, mechanism to provide model for the IoT is just in its infancy. One point
adequate reliability and error control with little of view, currently supported by some researchers,
IoT is even wider, overhead. Recent results have shown how these is to have a narrow waist above the network layer
we believe that paradigms can greatly improve the performance of
dissemination in homogeneous networks, but
and just below the service layer. By doing so, the
capabilities of the underlying IoT technologies
building an architec- extension of these techniques to highly heteroge- could be exposed as a universal building block.
neous scenarios has not yet been addressed. Final- Standardized interfaces could provide consistent
ture that is centered ly, for densely deployed nodes with very limited service primitives for applications and services.
individual capabilities it makes sense to look into While we can concede that this is a useful start-
on a similar concept distributed processing paradigms for decoding. ing point for a discussion, we strongly believe
•Clustering and cooperation: leveraging on that this issue needs to be investigated more
will be essential for the presence of more capable nodes to help oth- deeply, before a final conclusion can be drawn as
its success. ers in reliably delivering their data is a good
approach in a heterogeneous environment; an
to where the “next” narrow waist will be, and
how it will look like. The use of different wireless
interesting open research problem in this context protocol stacks for different applications and
is how to place a number of such nodes in an environments makes the choice of the “next”
environment so as to improve the overall net- narrow waist extremely complex. The conver-
work performance. Cooperation strategies, rang- gence between different communication mecha-
ing from PHY cooperation to improve the nisms into a coherent and interoperable fabric
quality of wireless links to opportunistic strate- might pass through the extension of some of the
gies at the MAC and routing layers, have been existing protocols, together with the creation of
shown to provide significantly better perfor- bridge routers, hosted in base stations.
mance in dynamic wireless networking contexts; As a first step towards drafting a convergent
we may expect that this approach will be even architecture for the wireless protocols in the IoT
more important in an IoT scenario, where het- area, we identify three main issues . We believe
erogeneity and scale are key factors. The chal- that the development of Machine to Machine
lenge in this case is how to provide simple yet interfaces (M2M API) will enable proactive
effective and robust distributed schemes. communication of devices, transparent to the
•Disruption-tolerant paradigms: techniques user. M2M APIs should be drawn up for all
to provide networking capabilities even in the devices of the identified application areas.
absence of stable connectivity will play an impor- Important efforts have been made in this area,
tant role in environments where continuous and groups such as the ETSI M2M Working
wireless coverage cannot be guaranteed and Group are developing interesting solutions. Fur-
where there may be a significant degree of ran- thermore, we think that a uniform protocol view,
domness in the communications relationships compatible with the current IP suite, will provide
among wireless nodes. Mechanisms to reveal the protocols at different levels and will be the basis
presence of other nodes and to decide whether of device interoperability. The development of
to exchange data will need to be developed, in interfaces, bridges and inter-protocol routers will
view of the fact that some nodes cannot afford allow all devices, generally developed with a pre-
to continuously transmit/listen but rather need cise service in mind, to embrace a greater variety
to judiciously manage their own energy. Scenar- of applications. Last but not least, a thorough
ios where mobile users (e.g., a person carrying a analysis of the effects of the IoT at the network
cell phone) can be used for data ferrying and/or level will be needed, paying particular attention
to provide information about, e.g., the network to the terabyte torrent generated by billions of
topology, seem very promising. interconnected devices and how it can be han-
dled via wireless communications.
A NEW NARROW WAIST BELOW
THE SERVICE LAYER LINKING PHYSICAL ENTITIES AND DEVICES
One of the main factors that contributed to the OF THE INTERNET OF THINGS
success of the current Internet architecture is the The IoT is expected to provide a resource fabric
centrality of the Internet Protocol (IP), which interfacing the physical world by means of a
can be seen as a narrow waist between connect- ubiquitously deployed substrate of embedded
ed devices on one side and applications and ser- networked devices. These resources provided by
vices on the other. This architectural choice the IoT include sensors, actuators, RFID tags

48 IEEE Wireless Communications • December 2010

ZORZI LAYOUT 12/9/10 11:15 AM Page 49

and readers, NFC enabled devices, etc. Some of

these devices will have the ability to interact with Resolution infrastructure
or record information/events concerning the real
world and the entities contained within. Real Monitoring of Discovery of entities Lookup of IoT
world entities can have living embodiments such associations and IoT resources
between entities resources id345 id123
as persons and animals or be objects such as and IoT resources id123
id345

buildings, cars, tools, production material or

Monitoring
Input
other organic and inorganic matter.
Business services, end user applications or
enabling middleware services can exploit the
information and interaction capabilities of the
IoT with respect to real world entities only if the
necessary set of available IoT devices can be
determined. This includes finding the relevant Moving IoT resource
entities and the corresponding IoT devices that
provide information about these entities or allow
interactions with them. Most of the existing
research has been based on the assumption that Moving entity
there is a static association between the
resources of the IoT and surrounding real world Fixed association
entities, or that associations can be inferred by Entity IoT Resource
globally available location information. In real Dynamic association
life the environment is much more heteroge-
neous and dynamic as IoT devices and real world
entities may be mobile with respect to each Figure 2. A new resolution infrastructure for linking physical entities and
other, IoT resource availability and communica- devices in the IoT.
tion capabilities may vary throughout an envi-
ronment and in time, and knowledge of locality
may be incomplete. Such an environment makes (WSNs) represent a reasonably cheap sensory
the resolution of available IoT devices that are extension to Internet-connected devices; more-
concerned with real world entities a challenging over, their computational capabilities allow for
task, particularly on a global scale. further (though possibly limited) flexibility of use
We envision the development of a new reso- and functional expansion. Any kind of next-gen-
lution infrastructure that enables an efficient eration Internet enabled portable device will set
linking of real world entities with relevant up advanced interactions with the “things” mak-
devices of the IoT. The provided functionality ing up the new IoT, resulting in a pervasive
consists of three different aspects which are visu- infrastructure of fixed and mobile heterogeneous
alized in Fig. 2, namely: nodes, seamlessly providing, exploiting or shar-
1. Discovery of the relevant entities, e.g., based ing context based services and applications. In
on identifier, location, type, provider, topic, or particular, capturing the context around the
a combination thereof devices will constitute a key component, making
2. Lookup of IoT devices that can provide infor- such operations as “Googling” the physical reali-
mation about the entities or allow interactions ty possible and common. By integrating any
with them object into the IP infrastructure, making it possi-
3. Monitoring IoT devices and entities and keep- ble to natively address and connect it, 6LoW-
ing the dynamic links between them up-to- PAN [14] is an important enabling technology to
date support the cited IoT interaction paradigm,
At the same time the developed resolution while still running everything over the
infrastructure has to ensure that only authorized widespread Internet infrastructure. The use of
services and applications can discover available Web services is a key ingredient of this migra-
devices of the IoT concerning real world entities. tion, and allows the transparent integration of
A particular challenge hereby is the handling of specialized systems (such as a WSN) with any
heterogeneous identification schemes that will system built with standard components, thereby
unavoidably exist in the Internet of Things. As greatly accelerating the penetration of the IoT
mentioned, earlier IoT devices can be identified paradigm. As an example, TinyREST is an effi-
by different identification schemes or communi- cient implementation of Web services in WSNs,
cation level identifiers. Similar heterogeneity can that accounts for the specific resource require-
be expected from the assignment of identities to ments of sensor nodes.
real world entities. The resolution infrastructure We recently put this vision into practice [13],
will be able to cope with this heterogeneity, by channeling experience in the field of wireless
allowing a controlled inter-linking of those based sensor networking towards the realization of a
on the privacy and security contexts of the IoT scalable and easily extendable network structure,
devices and their respective real world entities. which includes different classes of nodes running
compatible code but providing different func-
ARCHITECTURE AND PROTOCOLS FOR THE INTERNET tions and carrying out different tasks. The net-
work spans several floors in three buildings at
OF THINGS: A CASE STUDY [13] the University of Padova, and includes high- as
The future Internet will strive to yield novel well as low-density areas (Fig. 3). The focus is
means of interaction with services, other users, not only on the setup and installation of the net-
and the environment. Wireless Sensor Networks work (although this is by itself interesting and

IEEE Wireless Communications • December 2010 49

ZORZI LAYOUT 12/9/10 11:15 AM Page 50

Figure 3. Topology of a portion of the testbed at the University of Padova, shown through a Google Maps-based application [15].

challenging), but rather on its use, development that have not been solved, including heterogene-
and structure optimization. To this end, all ity, scalability, security, connectivity, energy,
nodes are IPv6-compatible, which makes them management, naming and identification. The
directly addressable from any Internet-capable complexity of these technical issues, especially in
device. The nodes support diverse services, from view of the resource-constrained nature of many
environmental parameter monitoring to localiza- IoT components and of the use of wireless com-
tion, and are in turn supported by lower-level munications, calls for a unified architectural view
functionalities allowing, e.g., to switch the appli- able to address them in a coherent fashion.
cation being run on the node, to change the After pointing out several of the main issues, we
class/role of a node, to spread software changes have provided some initial ideas to address some
and updates over the air, or to perform low-level of them. We have also briefly described a recent
resets in case of malfunctions. Offering services case study where an architecture and protocol
through our network provides an opportunity to suite for IoT has been implemented. These con-
realize part of the IoT vision and continue tributions, far from representing complete and
research efforts in the field. In addition, it con- definitive answers, are nevertheless a useful
cretely demonstrates the advantages of the IoT starting point for a deeper discussion that will
in the everyday management of a campus envi- keep researchers busy for the next decade.
ronment. A major research effort, IoT-A (Internet of
Experience with such a large network testbed, Things — Architecture, www.iot-a.eu), which
in which the above protocol suite has been fully represents the flagship project in the IoT area
implemented and tested, has shown that the IoT within the European Commission’s FP7, and
architecture based on Web services and which includes the authors among its key partici-
IPv6/6LoWPAN is a powerful paradigm with the pants, has started in September 2010 and will be
potential of allowing a rapid penetration of IoT running for three years.
concepts in practical systems. Details of the
implementation and additional discussion can be ACKNOWLEDGMENT
found in [13], whereas some experimental results This work has been partially supported by the
related to the use of such architecture in a smart European Commission under the IoT-A project
monitoring and smart grid scenario can be found in FP7 (Grant Agreement no. 257521). The
in [15]. authors would also like to thank the partners of
the IoT-A consortium, whose discussions have
CONCLUSIONS inspired some of the ideas presented in this arti-
cle.
In this article we have provided a brief descrip-
tion of the current state of the art of the Inter- REFERENCES
net of Things, with special focus on concepts and [1] X-ETP Group, Future Internet Strategic Research Agen-
technologies related to mobility and wireless da, Feb. 2010.
[2] L. Atzori, A. Iera, and G. Morabito, The Internet of
communications and networking. Although the Things: A Survey, Elsevier Computer Networks, 2010.
IoT concept has been around for several years [3] ITU Internet Reports, The Internet of Things (Ed. 2005),
now, there are still many major technical issues Nov. 2005.

50 IEEE Wireless Communications • December 2010

ZORZI LAYOUT 12/9/10 11:15 AM Page 51

[4] SENSEI FP7 Project, Scenario Portfolio, User and Context rial Board of the Wiley Journal of Wireless Communications
Requirements, Deliverable 1.1, Available at and Mobile Computing. He was also a guest editor for spe-
http://www.sensei-project.eu/. cial issues in IEEE Personal Communications (Energy Man-
[5] E. Welbourne et al., “Building the Internet of Things agement in Personal Communications Systems) and IEEE
Using RFID: The RFID Ecosystem Experience,” IEEE Inter- Journal on Selected Areas in Communications (Multimedia
net Computing, vol. 13, no. 3, May–June 2009, pp. Network Radios, Underwater Wireless Communication Net-
48–55. works). He is currently serving as a voting member of the
[6] A. Dada and F. Thiesse, “Sensor Applications in the ComSoc Board of Governors.
Supply Chain: The Example of Quality-Based Issuing of
Perishables,” Proc. Internet of Things 2008, Zurich, ALEX GLUHAK (A.Gluhak@surrey.ac.uk) is a senior researcher
Switzerland, May 2008. at the Centre for Communication System Research (CCSR)
[7] K. Finkenzeller, RFID Handbook, 3rd Ed., Wiley, 2010, at the University of Surrey, UK, where he is coordinating
ISBN: 978-0-470-69506-7. Internet of Things related research activities. He is technical
[8] D. Kofman, EPC Global Standards: An Opportunity or A manager of the ICT-FP7 SENSEI project, a European project
Threat for Europe, The Internet of the Future, Bled with 20 partners, investigating the integration of the physi-
2008. cal world into the Future Internet. He has held research
[9] B. Krishnamachari, Networking Wireless Sensors, Cam- positions with CSSR and later the Ericsson Ireland Research
bridge University Press, Jan. 2006, ISBN: 978-0-5218- Centre, while contributing actively to several large Euro-
3847-4 pean research projects, such as e-SENSE, SENSEI and
[10] C. Floerkemeier, R. Bhattacharyya, and S. Sarma, recently SmartSantander and IoT-A. His main research
“Beyond the ID in RFID,” Proc. TIWDC 2009, Pula, Italy, interests are next generation network architectures and
Sept. 2009. experimental facilities - in particular integration of the
[11] ETSI Machine to Machine Standardization Technical Internet of Things into Internet service layers, service-ori-
Committee, www.etsi.org/Website/Technologies/ ented sensor networks and scalable context information
M2M.aspx. infrastructures for next generation networks. He has been
[12] K. Sakamura, “Challenges in the Age of Ubiquitous visiting researcher at NICT Japan, in 2005 and University of
Computing: A Case Study of T-Engine — An Open California Irvine, US in 2002.
Development Platform for Embedded Systems,” Proc.
ICSE’06, Shanghai, China, May 2006. S EBASTIAN L ANGE (slange@vdivde-it.de) holds a degree in
[13] A. P. Castellani et al., “Architecture and Protocols for physics from the University of Heidelberg, Germany. After
the Internet of Things: A Case Study,” Proc. 1st IEEE his Ph.D. at the European Molecular Biology Laboratory in
Int’l. Wksp. Web of Things (WoT 2010 at IEEE PER- Heidelberg, Germany he has been working as management
COM), pp. 678–83. consultant for the Consulting Company Droege &
[14] Z. Shelby and C. Borman, 6LoWPAN: The Wireless Comp./Arideon with a focus on business-process and
Embedded Internet, Wiley, 2009, ISBN: 978-0-470- knowledge management. He has been working with
74799-5. VDI/VDE-IT since 2006 and is currently Senior Consultant in
[15] N. Bressan et al., “The Deployment of A Smart Moni- the Department Innovation Europe. He helped to establish
toring System Using Wireless Sensor Network and Actu- the European Technology Platform on Smart Systems Inte-
ator Networks,” Proc. IEEE SmartGridComm, gration where he is currently responsible for the manage-
Gaithersburg, MD, Oct. 4–6, 2010. ment of the ETPs Office as deputy secretary general and
holds conceptual and advisory functions. He has been
BIOGRAPHIES involved in the management of several EU FP projects and
MICHELE ZORZI [F’07] (zorzi@dei.unipd.it) received his Laurea has been strongly committed to establishing and evolving
and Ph.D. degrees in electrical engineering from the Uni- the topic of IoT on a European level in the recent years.
versity of Padova, in 1990 and 1994, respectively. During
the academic year 1992-1993, he was on leave at the Uni- ALESSANDRO BASSI (abassi@eecs.utk.edu) graduated from the
versity of California, San Diego (UCSD), attending graduate University of Milan in 1994, with soft computing and soft-
courses and doing research on multiple access in mobile ware engineering as majors. He joined Amadeus in 1997,
radio networks. In 1993 he joined the faculty of the Dipar- and moved to the University of Tennessee in 2000, where
timento di Elettronica e Informazione, Politecnico di he was involved in the seminal work of the Internet Back-
Milano, Italy. After spending three years with the Center plane Protocol. He then held a Research Visitor position at
for Wireless Communications at UCSD, in 1998 he joined the ENS in Lyon, France, where he developed the relation-
the School of Engineering of the University of Ferrara, Italy, ship between storage and active networking. After working
and in 2003 joined the Department of Information Engi- for RIPE NCC, in November 2004 he joined Hitachi. He was
neering of the University of Padova, Italy, where he is cur- the deputy project coordinator of the FP5 6QM project,
rently a professor. His present research interests include and the project coordinator of the FP7 “Autonomic Inter-
performance evaluation in mobile communications sys- net” project. He also participated to several other EU pro-
tems, random access in mobile radio networks, ad hoc and jects and initiatives. Since 2007, his research interests are
sensor networks, energy constrained communications pro- focused on RFID technology and the Internet of Things. He
tocols, cognitive radios and networks, and underwater is the chair of the Internet of Things WG of the European
acoustic communications and networking. He was Editor- Technology Platform EPoSS. He is an expert for ENISA on
in-Chief of the IEEE Wireless Communications Magazine possible threats from IoT technology adoption, and he is
from 2003 to 2005, is currently Editor-In-Chief of the IEEE the Technical Coordinator of the IP FP7 project “Internet-
Transactions on Communications, and serves on the Edito- of-Things Architecture” (IoT-A).

IEEE Wireless Communications • December 2010 51