Internet of Things - Vision, Applications and Research Challenges

Ad Hoc Networks 10 (2012) 14971516
Contents lists available at SciVerse ScienceDirect
Ad Hoc Networks
journal homepage: www.elsevier.com/locate/adhoc
Survey Paper
Internet of things: Vision, applications and research challenges

Daniele Miorandi a,, Sabrina Sicari b, Francesco De Pellegrini a, Imrich Chlamtac a
a
CREATE-NET, via Alla Cascata 56/D, IT-38123 Povo, Trento, Italy
b
Dipartimento di Informatica e Comunicazione, Universit degli Studi dell Insubria, via Mazzini, 5, IT-21100 Varese, Italy
a r t i c l e i n f o a b s t r a c t
Article history: The term Internet-of-Things is used as an umbrella keyword for covering various aspects
Received 17 February 2012 related to the extension of the Internet and the Web into the physical realm, by means of
Accepted 25 February 2012 the widespread deployment of spatially distributed devices with embedded identication,
Available online 21 April 2012
sensing and/or actuation capabilities. Internet-of-Things envisions a future in which digital
and physical entities can be linked, by means of appropriate information and communica-
Keywords: tion technologies, to enable a whole new class of applications and services. In this article,
Internet-of-Things
we present a survey of technologies, applications and research challenges for Internet-
Web
Smart objects
of-Things.
RFID 2012 Elsevier B.V. All rights reserved.
Sensors
Actuators
Interoperability
Security
1. Introduction In such a perspective, the conventional concept of the

Internet as an infrastructure network reaching out to
Nowadays, around two billions people around the end-users terminals will fade, leaving space to a notion
world use the Internet for browsing the Web, sending of interconnected smart objects forming pervasive com-
and receiving emails, accessing multimedia content and puting environments [1]. The Internet infrastructure will
services, playing games, using social networking applica- not disappear. On the contrary, it will retain its vital role
tions and many other tasks. While more and more people as global backbone for worldwide information sharing
will gain access to such a global information and commu- and diffusion, interconnecting physical objects with com-
nication infrastructure, another big leap forward is coming, puting/communication capabilities across a wide range of
related to the use of the Internet as a global platform for services and technologies.
letting machines and smart objects communicate, dia- This innovation will be enabled by the embedding of
logue, compute and coordinate. electronics into everyday physical objects, making them
It is predictable that, within the next decade, the Inter- smart and letting them seamlessly integrate within the
net will exist as a seamless fabric of classic networks and global resulting cyberphysical infrastructure. This will give
networked objects. Content and services will be all around rise to new opportunities for the Information and Commu-
us, always available, paving the way to new applications, nication Technologies (ICT) sector, paving the way to new
enabling new ways of working; new ways of interacting; services and applications able to leverage the interconnec-
new ways of entertainment; new ways of living. tion of physical and virtual realms.
Within such perspective, the term Internet-of-Things
(IoT) is broadly used to refer to both: (i) the resulting
Corresponding author. Tel.: +39 0461 40 84 00; fax: +39 0461 42 11 57. global network interconnecting smart objects by means
E-mail addresses: [email protected] (D. Miorandi),
of extended Internet technologies, (ii) the set of supporting
[email protected] (S. Sicari), francesco.depellegrini@create-net.
org (F. De Pellegrini), [email protected] (I. Chlamtac). technologies necessary to realize such a vision (including,
1570-8705/$ - see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.adhoc.2012.02.016
1498 D. Miorandi et al. / Ad Hoc Networks 10 (2012) 14971516
e.g., RFIDs, sensor/actuators, machine-to-machine commu- main (hosts, terminals, routers, etc.) [7]. We dene smart
nication devices, etc.) and (iii) the ensemble of applications objects (or things) as entities that:
and services leveraging such technologies to open new
business and market opportunities [2,3]. Have a physical embodiment and a set of associated
In this survey article, we aim at providing a holistic per- physical features (e.g., size, shape, etc.).
spective on the Internet-of-Things concept and develop- Have a minimal set of communication functionalities,
ment, including a critical revision of application elds, such as the ability to be discovered and to accept
enabling technologies and research challenges. As a matter incoming messages and reply to them.
of fact, the research community active on IoT-related Possess a unique identier.
themes is still highly fragmented, and, to a large extent, Are associated to at least one name and one address.
focused around single application domains or single tech- The name is a human-readable description of the object
nologies. Further, the involvement of the networking and and can be used for reasoning purposes. The address is a
communications scientic communities is still limited, de- machine-readable string that can be used to communi-
spite the high potential impact of their contributions on cate to the object.1
the development of the eld [2,4]. We do believe that this Possess some basic computing capabilities. This can
fragmentation is potentially harmful for the development range from the ability to match an incoming message
and successful adoption of IoT technologies. We therefore to a given footprint (as in passive RFIDs) to the ability
hope this survey can help in bridging existing communi- of performing rather complex computations, including
ties, fostering cross-collaborations and ensuring that service discovery and network management tasks.
IoT-related challenges are tackled within a system-level May possess means to sense physical phenomena (e.g.,
perspective, ensuring that the research activities can then temperature, light, electromagnetic radiation level) or
be turned into successful innovation and industry to trigger actions having an effect on the physical reality
exploitation. (actuators).
The remainder of this article is organized as follows. In
Section 2 we introduce the IoT vision and dene the main The last point in the denition above is the key one, and
related concepts. In Section 3 we analyze the relevant re- differentiates smart objects from entities traditionally con-
search and technology contexts, including related elds sidered in networked systems. In particular, the proposed
and their potential contribution towards the realization classication includes devices considered in RFID research
of the IoT vision. In Section 4 we present the main research [8] as well as those considered in wireless sensor networks
challenges ahead of us in the IoT landscape. In Section 5 we (WSNs) and sensor/actor networks (SANETs) [9,10].
discuss the security challenges introduced by IoT technol- The inclusion of such entities into a global networked
ogies and applications. An analysis of the potential applica- system questions the architectural and algorithmic princi-
tion elds and impact areas is reported in Section 6. A ples at the basis of the design of the Internet as we know it.
survey of IoT related-on-going initiatives is presented in In particular, the increased level of heterogeneity, due to
Section 7. Section 8 concludes the survey with a number the inclusion of devices with only very basic communica-
of remarks on potential approaches to tackle the chal- tion and computing capabilities, challenges the assump-
lenges identied. tion that any device presents a full protocol stack, as well
as the application of the end-to-end principle in network
operations [11]. From the conceptual standpoint, indeed,
2. Vision and concept IoT is about entities acting as providers and/or consumers
of data related to the physical world. The focus is on data
The Internet-of-Things is emerging as one of the major and information rather than on point-to-point communi-
trends shaping the development of technologies in the ICT cations. This fact could push towards the adoption of re-
sector at large [3,5,6,2]. The shift from an Internet used for cently proposed content-centric network architectures
interconnecting end-user devices to an Internet used for and principles [12], as will be discussed in the following
interconnecting physical objects that communicate with sections.
each other and/or with humans in order to offer a given From a system-level perspective, the Internet-of-Things
service encompasses the need to rethink anew some of can be looked at as a highly dynamic and radically distrib-
the conventional approaches customarily used in network- uted networked system, composed of a very large number
ing, computing and service provisioning/management. of smart objects producing and consuming information.
From a conceptual standpoint, the IoT builds on three The ability to interface with the physical realm is achieved
pillars, related to the ability of smart objects to: (i) be iden- through the presence of devices able to sense physical phe-
tiable (anything identies itself), (ii) to communicate (any- nomena and translate them into a stream of information
thing communicates) and (iii) to interact (anything interacts) data (thereby providing information on the current context
either among themselves, building networks of intercon- and/or environment), as well as through the presence of
nected objects, or with end-users or other entities in the devices able to trigger actions having an impact on the
network. Developing technologies and solutions for en- physical realm (through suitable actuators). As scalability
abling such a vision is the main challenge ahead of us. is expected to become a major issue due to the extremely
At the single component level, the IoT will be based on
the notion of smart objects, or, simply, things, which 1
Their association and relation to the identier will be discussed further
will complement the existing entities in the Internet do- later on in this article.
D. Miorandi et al. / Ad Hoc Networks 10 (2012) 14971516 1499
large scale of the resulting system, and considering also the tectural and protocol levels. In particular, this may
high level of dynamism in the network (as smart objects question the thin waist approach at the basis of IP
can move and create ad hoc connections with nearby ones networking.
following unpredictable patterns), the quest for inclusion Scalability. As everyday objects get connected to a global
of self-management and autonomic capabilities is ex- information infrastructure, scalability issues arise at
pected to become a major driver in the development of a different levels, including: (i) naming and addressing
set of enabling solutions [13,14]. due to the sheer size of the resulting system, (ii) data
From a service-level perspective, the main issue relate communication and networking due to the high level
to how to integrate (or: compose) the functionalities and/ of interconnection among a large number of entities,
or resources provided by smart objects (in many cases in (iii) information and knowledge management due to
forms of data streams generated) into services [1517]. the possibility of building a digital counterpart to any
This requires the denition of: (i) architectures and meth- entity and/or phenomena in the physical realm and
ods for virtualizing objects by creating a standardized (iv) service provisioning and management due to
representation of smart objects in the digital domain, able the massive number of services/service execution
to hinder the heterogeneity of devices/resources and (ii) options that could be available and the need to handle
methods for seamlessly integrating and composing the re- heterogeneous resources.
sources/services of smart objects into value-added services Ubiquitous data exchange through proximity wireless
for end users. technologies. In IoT, a prominent role will be played by
The Internet-of-Things vision provides a large set of wireless communications technologies, which will
opportunities to users, manufacturers and companies. In enable smart objects to become networked. The ubiqui-
fact, IoT technologies will nd wide applicability in many tous adoption of the wireless medium for exchanging
productive sectors including, e.g., environmental monitor- data may pose issues in terms of spectrum availability,
ing, health-care, inventory and product management, pushing towards the adoption of cognitive/dynamic
workplace and home support, security and surveillance radio systems [20].
(see Section 6 for a more in-depth discussion of relevant Energy-optimized solutions. For a variety of IoT entities,
application domains). minimizing the energy to be spent for communica-
From a user point of view, the IoT will enable a large tion/computing purposes will be a primary constraint.
amount of new always responsive services, which shall an- While techniques related to energy harvesting (by
swer to users needs and support them in everyday activi- means, e.g., of piezolectric materials or micro solar pan-
ties. The arising of IoT will provide a shift in service els) will relieve devices from the constraints imposed
provisioning, moving from the current vision of always- by battery operations, energy will always be a scarce
on services, typical of the Web era, to always-responsive sit- resource to be handled with care. Thereby the need to
uated services, built and composed at run-time to respond devise solutions that tend to optimize energy usage
to a specic need and able to account for the users context. (even at the expenses of performance) will become
When a user has specic needs, she will make a request more and more attractive.
and an ad hoc application, automatically composed and Localization and tracking capabilities. As entities in IoT
deployed at run-time and tailored to the specic context can be identied and are provided with short-range
the user is in, will satisfy them. wireless communications capabilities, it becomes possi-
While the IoT vision will require substantial advances in ble to track the location (and the movement) of smart
a number of ICT elds (see Section 4), its realization is objects in the physical realm. This is particularly impor-
likely going to follow an incremental process, starting from tant for application in logistics and product life-cycle
existing technologies and applications. In particular, IoT management, which are already extensively adopting
will likely expand starting from identication technologies RFID technologies.
such as RFID (Radio Frequency Identication) [8,18], which Self-organization capabilities. The complexity and
are already widely used in a number of applications. At the dynamics that many IoT scenarios will likely present
same time, in its development path, IoT will likely build on calls for distributing intelligence in the system, making
approaches introduced in a variety of relevant eld, such as smart objects (or a subset thereof) able to autono-
wireless sensor networks (as a means to collect contextual mously react to a wide range of different situations, in
data [9]) and service-oriented architectures (SoA) as the order to minimize human intervention. Following users
software architectural approach for expanding Web-based requests, nodes in IoT will organize themselves autono-
services through IoT capabilities [19]. mously into transient ad hoc networks, providing the
Summarizing, we can preliminarily identify the follow- basic means for sharing data and for performing coordi-
ing key system-level features that Internet-of-Things nated tasks [21]. This includes ability to perform device
needs to support: and service discovery without requiring an external
trigger, to build overlays and to adaptively tune proto-
Devices heterogeneity. IoT will be characterized by a cols behavior to adapt to the current context [13].
large heterogeneity in terms of devices taking part in Semantic interoperability and data management. IoT will
the system, which are expected to present very differ- be much about exchanging and analyzing massive
ent capabilities from the computational and communi- amounts of data. In order to turn them into useful infor-
cation standpoints. The management of such a high mation and to ensure interoperability among different
level of heterogeneity shall be supported at both archi- applications, it is necessary to provide data with
adequate and standardized formats, models and seman- common characteristics, and to share a number of common
tic description of their content (meta-data), using well- features including a full protocol stack. While advances in
dened languages and formats. This will enable IoT embedded electronics and software are making such a
applications to support automated reasoning, a key fea- requirement less and less stringent [23], it still appears
ture for enabling the successful adoption of such a tech- to put unnecessary burden on the devices. Solutions able
nology on a wide scale. to accommodate heterogeneity in terms of supported fea-
Embedded security and privacy-preserving mechanisms. tures should be introduced to ease incremental deploy-
Due to the tight entanglement with the physical realm, ment. The second factor relates to the need of equipping
IoT technology should be secure and privacy-preserving sensor nodes with a battery. While a number of solutions
by design. This means that security should be consid- for increasing energy efciency at various layers of the
ered a key system-level property, and be taken into OSI model has been devised, the need to replace batteries
account in the design of architectures and methods for from time to time represents a huge barrier to the wide-
IoT solutions. This is expected to represent a key spread development of such technology. A number of
requirements for ensuring acceptance by users and promising research lines, related to energy harvesting
the wide adoption of the technology. [24] and passive wireless sensor networks [25] are cur-
rently under development. The third issue relates to the
3. Research context dimension of the electronics needed to be embedded in ob-
jects to make them part of the IoT world. While recent ad-
As technology progresses, more and more processing vances in microelectronics have led to considerable
power, storage and battery capacity become available at reduction in size, the current state-of-the-art is unlikely
relatively low cost and with limited space requirements. to be sufcient to enable the realization of the full IoT vi-
This trend is enabling the development of extremely sion. In this respect, applications of nanotechnologies,
small-scale electronic devices with identication/commu- while still in their infancy [26], may represent a promising
nication/computing capabilities, which could be embedded research direction for extending the scope and applicabil-
in the environment or in common objects. Such a class of ity of IoT solutions.
devices could be used, as described in the previous Section, Radio frequency identication devices and solutions
to enable a set of novel applications and services, leverag- can nowadays be considered a mainstream communica-
ing direct interactions with the physical realm. The devel- tion technology, with a number of massive deployments,
opment of such a new class of services will, in turn, require in particular in the goods management and logistics sec-
the introduction of novel paradigms and solutions for com- tors. RFID is expected to play a key role as enabling identi-
munications, networking, computing and software engi- cation technology in IoT. At the same time, its integration
neering. The IoT umbrella concept comprises all these with sensing technologies brings alongside a number of
aspects, based on the paradigm of computing and commu- challenges and issues [27,28]. RFID applications have been
nications anywhere, anytime and by anything. so far mainly thought for use within isolated, vertically
In this section, we briey discuss the relevance and po- integrated, systems, used only for identication and/or
tential impact of existing research areas on the develop- tracking of objects embedded with an RFID tag. Their use
ment of IoT technologies and applications. as part of a larger system, where identication of an object
The Internet-of-Things is unlikely to arise as a brand is only a step of the work-ow to be executed to provide a
new class of systems. We envision an incremental develop- nal service, has not been fully explored yet.
ment path, along which IoT technologies will be progres- IoT shares a number of characteristics with ambient
sively employed to extend existing ICT systems/ intelligence [29]. In Ambient Intelligence (AmI), environ-
applications, providing additional functionalities related ments rich in sensing/computing/actuation capabilities
to the ability of interacting with the physical realm. In this are designed so to respond in an intelligent way to the
sense, we do believe it is worth analyzing which research presence of users, thereby supporting them in carrying
elds, among the ones subject of investigation in the last out specic tasks. Ambient intelligence builds upon the
years, can be more relevant (in terms of techniques/solu- ubiquitous computing concept, loosely dened as the
tions introduced or lessons learned) in the IoT scenario. embedding of computational devices into the environ-
In terms of enabling technologies, a key issue for IoT is ment. Ubiquitous computing provides therefore the dis-
the development of appropriate means for identifying tributed infrastructure necessary to enable the
smart objects and enabling interactions with the environ- development of AmI applications.
ment. In this sense, key building blocks are expected to AmI shares with IoT a number of aspects. This com-
be represented by wireless sensor networking technologies prises the inclusion in the system of sensing/computing
[9] and RFID [8,18,22]. capabilities embedded in the environment. At the same
As far as wireless sensor nodes and networks are con- time, AmI applications have been mainly developed for
cerned, the ability of sensing the environment and to closed environments (e.g., a room, a building), whereby
self-organize into ad hoc networks represent important a number of specic functions (known at design time)
features from an IoT perspective. At the same time, three can be accommodated and supported. Accordingly, one of
main limiting factors need to be overcome in order to fos- the main focus of research in AmI has been the develop-
ter their widespread adoption. The rst one relates to the ment of reasoning techniques for inferring activities of
support of heterogeneous devices. Nodes in a wireless sen- users and devising appropriate response strategies from
sor network are customarily expected to possess a set of the embedded devices. IoT expands the AmI concepts to
integrate open scenarios, whereby new functions/capa- Further, most of the approaches proposed have been con-
bilities/services need to be accommodated at run-time ceived to be applied to a single, well-dened specic
without them having been necessarily considered at de- application eld. What is needed to foster the deployment
sign time. This requires IoT solutions to be inherently of IoT applications is instead a set of design patterns that
autonomic, i.e., presenting the self-conguration and self- can be used to augment end-user applications with self-
organization, possibly cognitive, capabilities needed to adaptive properties. This requires methods for discover-
provide this additional degree of exibility. ing, deploying and composing services at run-time in a
IoT application scenarios require applications to distributed fashion, supporting autonomicity within all
prove adaptable to highly diverse contexts, with differ- phases of the service life-cycle. While smart objects may
ent resources available and possibly deployment envi- be able to run some limited and lightweight services,
ronments changing over time. A number of approaches one key aspect of IoT is the integration with the Internet
have been proposed to overcome devices heterogeneity infrastructure, i.e., the cloud. This may take the form of
in related scenarios. In particular, the use of a standard appropriate Web-based services and applications, able to
virtual platform in all network devices has been pro- leverage data and/or atomic services made available by
posed [30,31]. While this approach has the potential smart things to provide value-added services to the end
to ease the development of software and services for user.
IoT by providing a standard set of supported primitives, As far as frameworks for developing IoT applications are
at the same time it poses some rather stringent require- concerned, a major role is expected to be played by ap-
ments on the hardware capabilities of the devices them- proaches based on so-called service-oriented computing
selves. Frameworks based on mobile agents have also (SOC) [4345]. SOC envisages a possibly distributed archi-
been proposed, e.g. [32]. Their applicability to IoT envi- tecture, whereby entities are treated in a uniform way and
ronments may however prove difcult due, again, to the accessed via standard interfaces. A service-oriented archi-
expected high level of heterogeneity in the resources tecture (SOA) is essentially a collection of services, which
available on devices. communicate with each other via a set of standardized
All the efforts required in terms of development of IoT interaction patterns. The communication can involve
architectures, methods for management of resources, dis- either simple message passing or it could involve two or
tributed communication and computation, represent the more services coordinating some activity via appropriate
baseline for the introduction of innovative services that protocols. Currently, many SOC deployments make use of
will improve users experience and quality of life. As de- Web-based protocols (e.g., http) for supporting interoper-
scribed in the previous section, IoT services will be respon- ability across administrative domains and enabling tech-
sive in nature, being able to anticipate user needs, nologies. SOC can be used to manage web services and
according to the situation they are in, by means of dynamic make them act like a virtual network, adapting applica-
resource management schemes and on-the-y composi- tions to the specic users needs. Service-oriented architec-
tion of different service components. tures support a given level of heterogeneity and exibility
This requires applications to be able to understand the in the software modules to be deployed and executed
context and situation the user is in. Such a theme has [44,43,46]. SOC/SOA in general and Web services in partic-
been addressed within the ambient intelligence, ambient ular cannot be straightforwardly applied to the construc-
assisted living and pervasive computing elds, leading tion of IoT applications. In particular, such approaches
to a number of solutions able to leverage contextual at least in their current form may prove too heavyweight
information coming from a number of sources. In [33] a for being deployed on resources-constrained devices.
contextual information service is introduced, which pro- Nonetheless, they represent a very powerful approach in
vides applications with contextual information via a vir- terms of abstracting functionality from the specic soft-
tual database in an efcient and scalable way. In this ware implementation as well as for ensuring integration
direction other solutions have been proposed for provid- and compatibility of IoT technologies into the bigger Fu-
ing applications with contextual information in a distrib- ture Internet-Future Web perspective, a key success factor
uted setting [3436]. Schilits active map system [34,35] for enabling the IoT vision. In particular, exploiting the po-
represents a location-based publish-subscribe system for tential of solutions based on Web service technology may
contextual information dissemination. In such a system, ease the development of a new exible, dynamic and open
location-tagged contextual information is published to platform of services for Internet-of-Things with a set of
an active map server, which disseminates the information self- methods for the distributed and autonomic manage-
to subscribed applications. Another approach is Easyliving ment and run-time optimization of the platform itself. Key
[36], which stores contextual information in a single data- concept from SOA/SOC, such as late binding and dynamic
base, allowing applications to query it in order to retrieve service composition/orchestration, are expected to be
data. inherited in IoT. At the same time, new methods are neces-
Services in IoT are expected to be able to seamlessly sary to adapt them to the IoT peculiarities, including the
adapt to different situations and contexts. A number of denition of specic data models and representation,
research efforts for building self-adaptive situated architectures and methods for virtualizing smart objects
services have been undertaken in the last few years and their services/resources, together with the develop-
[3742]. However, we are still far from reaching a global ment of new methods for the dynamic and exible compo-
understanding of how to develop self-adaptive services sition of smart objects into the Internet of Services
presenting the exibility level required by IoT scenarios. [47,16,17].
4. Research challenges 3. Anything interacts: smart things can interact with the
local environment through sensing and actuation capa-
The key idea behind the Internet-of-Things concept, as bilities whenever present.
outlined in Section 2, resides in the huge potential of
embedding computing and communication capabilities Based on the aforementioned considerations, in the fol-
into objects of common use. Two additional features lowing we make an attempt to classify the research chal-
should also be properly accounted for: lenges that need to be addressed in order to turn the
Internet-of-Things from a concept into a well engineered,
Identication. Each object should be identiable. commercially viable technological paradigm.
Depending on the specic scenarios, objects may
require to be uniquely identied, or to be identied as 4.1. Computing, communication and identication
belonging to a given class (e.g., this object is a pen, technologies
regardless of which pen it is). This could be done basi-
cally in two ways. The rst one is to physically tag The scenarios envisioned for IoT require the develop-
one object by means of RFIDs, QR code or similar. In ment of advanced techniques able to embed computing,
such a way an object can be read by means of an communication and identication capabilities into every-
appropriate device, returning an identier that can be day objects. In the last years, several aspects have been
looked up in a database for retrieving the set of features investigated in related elds. The span is wide, ranging
(description) associated to it. The second possibility is from the research on low-cost low-power consumption
to provide one object with its own description: if micro/nano-electronics (for both computing as well as
equipped with wireless communication means, it could communication purposes), to advancement in near-eld
communicate directly its own identity and relevant fea- communications (RFID-like) for identication purposes.
tures. These two approaches are not mutually exclusive, Low-power communications is a well-established re-
and can complement each other. RFID-based identica- search eld within the sensor networking community, as
tion is indeed cheaper in terms of requirements on the proved by the active research performed in the last decade
electronics to be embedded in objects, but requires the on power consumption aware medium access protocols
possibility for the reader to access a database where [4852]. The typical approach pursued in such works re-
information about such an object is stored. The self- lates to the match of the RF front-end activation patterns
description-based approach, on the contrary, relaxes (i.e., sleep periods) to the trafc pattern. The use of such
the requirements to access to a global database, but still protocols, however, at present does not provide a nal an-
requires to embed more electronics into everyday swer to the optimization of energy consumption versus
objects. scalability issues. These are of paramount importance for
Sensing/Actuation. Objects can interface with the physi- IoT scenarios, as battery replacement is a costly process
cal environment either passively, i.e., performing sens- to be avoided as much as possible, especially for large-
ing operations, or actively, i.e., performing actions. scale deployments. Furthermore, the basic idea of such
These two dimensions span the two fundamental oper- protocols is to perform active/sleep duty cycles in order
ations that represent the interface and the coupling to save the power dispersed in idle listening. The increase
between the digital and the physical realms. Sensor/ in message latency [48] in turn needs to be traded off in or-
actor networks (SANETs) [10] have represented an der to balance between network lifetime and communica-
active research eld over the last decade. However, they tion performance.
have been mostly intended as ad hoc systems, with lim- More recently, advances in the eld of nano-scale accu-
ited physical extension and designed to carry out typi- mulators as well as energy harvesting techniques appear of
cally a single task. On the other hand, the IoT vision prominent interest to limit the need for battery replace-
requires to extend such a perspective considerably ments. In particular, it has been showed that it is possible
beyond current state-of-art technology. The main dif- to integrate several sources of energy harvesting into sen-
ference is that objects themselves could embed means sors, including piezoelectric, thermoelectric and radio
for sensing the local environments and acting on it, waves recharging devices [53]. A comprehensive take at
without being a priori bound to a single task/ the technological problem of energy harvesting in real de-
application. vices is described in [54]. There, techniques for power
management with the adaptation of sensor duty cycles
We can briey resume the three main system-level are proposed.
characteristics of the Internet-of-Things as follows: The effort to reduce the speed of discarding of IoT de-
vices has another dimension of particular relevance, which
1. Anything communicates: smart things have the ability to relates to the reciprocal interaction between computation
wirelessly communicate among themselves, and form and communication. The notion of distributing computa-
ad hoc networks of interconnected objects. tion in order to reduce the communication overhead,
2. Anything is identied: smart things are identied with a which is generally termed in-network processing or in-
digital name: relationships among things can be speci- network computing [55], is typically applied to wireless
ed in the digital domain whenever physical intercon- sensor networks that perform local measurements, as it
nection cannot be established. would be the case of eld measurements in IoT scenarios.
There, the natural requirement (and also the concern) is to activities, research on localization systems has tackled a
scale to a large number of sensor nodes. In order to in- number of issues that are certainly relevant in IoT research.
crease scalability, following the seminal work of Gupta One topic addressed involves the surveillance of moving
and Kumar on the scaling of capacity in wireless networks objects within a sensorized area [68], or the robustness
[56], several schemes for distributed estimation based on of location detection schemes [69] as needed in the
local communications have been proposed. For example, case of emergency networks [70]. Also, the recent advance-
authors of [57] proved that the best linear unbiased esti- ment of ultra-wideband radio frequency technologies
mation of a deterministic parameter can be computed at stimulated research for very ne-grained location estima-
every sensor with a distributed algorithm. Similarly, the tion and ranging [71].
scheme from [58] produces an estimate of the average va- Identication and proximity detection schemes that
lue of a random eld at each sensor. Average eld mea- make use of inexpensive RFIDs became recently a promis-
surement is performed by the distributed self-clocking ing choice for commercial deployments in the logistics
scheme described in [59]. Other approaches combine pack- eld [72]. The most popular type of RFIDs are passive tags,
et forwarding and computation as in [60], which uses a which do not contain an on-board power source: energy
combination of a binary split-tree algorithm coupled to a for operation is supplied by the RFID interrogation signal
binary hypothesis testing procedure. A joint MAC/PHY de- itself. Conversely, active tags have an on-board power
sign is proposed in [61], proving an asymptotically optimal source that feeds the on-board receiver and transmitter,
MAC for type-based estimation. In literature, the seminal allowing for an increased radio range. Semi-active and
paper exploring the issue is the Gallaghers scheme [62], semi-passive RFIDs differ in that the on-board power
where, under the assumption of perfectly scheduled com- source is used to feed the microchip, whereas transmission
munication, the proposed solution would permit the parity is either active (semi-active) or performed using back-scat-
check on the binary status of a set of nodes with required tering (semi-passive). Several vendors propose proprietary
communication complexity O(log log n). The later work in middleware platforms that have been developed with the
[63] proved that, in the case of type-threshold functions, aim to support commercial deployments of RFIDs; see for
such as AND, OR and MAJORITY, computing requires O(n) example the SAP Auto-ID Infrastructure [73]. Other plat-
broadcasts. Recently, the problem has been addressed by form include the Siemens RFID Middleware, Sun Java Sys-
the works in [55,64], proving fundamental scaling laws in tem RFID Software or the IBM WebSphere RFID.
the case of co-located and multi-hop packet networks. Ultimately, the main challenge from the communica-
The works [55,64] prove that there exist a strong depen- tion/computing perspective that hides behind the IoT con-
dence on the scaling law of the number of messages ex- cept is the need for an architecture supporting low-power,
changed and the computed function. low-cost and yet fully networked and integrated devices fully
Clearly, scaling issues arise when the need is either to compatible with standard communication technologies.
cover large areas with a grid of small-size devices for sens-
ing purposes, or to deploy a very dense one for localized 4.2. Distributed systems technology
measurements. Both cases apply indeed to IoT scenarios:
how to reconcile scaling laws derived in the context of in-net- This area includes all aspects related to enabling objects
work computing and ad hoc communications with a practi- to build a network, creating a distributed platform that en-
cally viable IoT architecture represents a major research able the easy implementation of services on top. This
challenge. builds on a traditional research line in computer science
At present the issues of density of deployed IoT devices [74,75], where a distributed system is dened as a system
are probably less critical, though. This relates to miniatur- driven by separate components which may be executed
ization of sensing and transmitting devices, a celebrated either sequentially or in parallel on different, intercon-
dimension of the research in sensing technologies which nected, nodes. The design of architectures and protocols
is not meeting the expectations set. Indeed, current tech- for distributed systems is a key issue for general net-
nologies are far from the level of integration foreseen in worked systems and for IoT in particular. In particular, sev-
the SmartDust vision [65]. The dimensions of commercial eral issues, involved in the design of IoT as a distributed
devices such as WASPMote2 or equivalent ones are typically system, can be identied. The analysis and design of IoT
of the order of 35 cm, dictated by the packaging dimension, cannot overlook aspects related to networking technolo-
mostly due to the RF interface dimensions and the volume of gies such as routing protocols, ow control robustness,
batteries. Nevertheless, notable advances have been made and synchronization. Problems like leader-election, node
with respect to the variety and the integration of sensing de- counting and averages computation are a core topic in
vices that are hosted on modern sensing boards: photocells the distributed systems literature [7678]. Part of such re-
for light measurements, thermistores for temperature search lines have been already re-discovered and renewed
probes, microphones, accelerometers and magnetometers in sensor networks literature, as recalled in the previous
represent standard equipment for modern sensor boards. section [79,80].
Localization systems represent a rather old research The distributed implementation of routing protocols is
line, dating back to early 90s, see for example Active Badge one of the fundamental algorithmic building blocks for
of Olivetti Research Ltd. [66] and Georgia Institute of networked systems [81]. However, as seen above, scalabil-
Technology CyberGuide [67]. Along some twenty years of ity issues discourage multi-hop communications for envi-
ronmental data retrieval, i.e., massive and large scale
2
http://www.libelium.com/products/waspmote. sensor networks do not appear a viable solution for IoT,
at present. Alternative architectures may make use of nate communications and computing in order to leverage
proximity communications whenever possible in case of the data coming from several information sources.
large deployments; possible implementations are de- IoT scenarios will be typically characterized by huge
scribed in [82,83]. amounts of data made available. A challenging task is to
The massive amount of data streaming from the envi- interpret such data and reason about it. This underpins
ronment to the Internet is a side effect of the IoT type of the need to have an actionable representation of IoT data
scenarios: this means a potentially very large amount of and data streams. This represents a key issue in order to
information injected into the network. The control of infor- achieve re-usability of components and services, together
mation injected by objects and related data ltering with interoperabilty among IoT solutions. Advances in data
techniques is a concern for pervasive scenarios [84]. Dis- mining and knowledge representation/management will
tributed ow control, in turn, is a well-studied traditional also be required, to satisfactorily address the peculiar fea-
topic in networking and controls due to the large amount tures of IoT technologies.
of work on TCP [85]. Surprisingly, how to control the huge A related research eld is that of distributed articial
amount of data injected into the network from the environ- intelligence, which addresses how autonomous software
ment is a problem so far mostly neglected in the IoT research. entities, usually referred to as agents, can be made able
Robustness and fault tolerance will become fundamental to interact with the environment and among themselves
topics in IoT scenarios, involving both the impact of com- in such a way to effectively pursue a given global goal
munication links failure, nodes software and hardware fail- [93]. Notice that in this domain a major challenge has to
ures, critical data integrity and general safety aspects. For a be faced, compared to the traditional design of a distrib-
general reference on fault tolerance and robustness ap- uted system. In fact, consider a simple task that involves
proaches please refer to [86]. Issues concerning the impact the coordination of several autonomous entities: e.g., vot-
of misbehaving nodes [87] represent also a traditional to- ing, auctioning, or cluster formation. The design of such
pic that is of interest for large-scale distributed systems applications has to account for the fact that part of the
as those foreseen in IoT. For a large-scale IoT deployment, control resides on single agents. Those are the entities that
the presence of myriads of devices in the environment requires ultimately interact and may choose different strategies
to replace/repair/reprogramm faulty, possibly embedded de- depending on a certain utility function. Thus, at system
vices and to design a system natively robust to failures of sin- design time, it is possible to leverage the theory of com-
gle nodes or groups thereof. petitive/cooperative games and let agents compete/form
Synchronization of clocks for tasks, which might under- coalitions upon their needs [94]. Theoretical foundations
go failures and restart has also been addressed extensively for these topics are rooted in game theory and social wel-
in the literature [88]. In the IoT scenario, the foreseen large fare. Applications to networking problems emerged only
scale enriches the challenges for both for data consistency recently. A technical description of the issues arising in
reasons and protocols functioning purposes. that context are beyond the scope of this survey: for a
From the implementation standpoint, a key issue is to standard reference see [95]. The access to the IoT devices
ease the inter-working from an application perspective. is unlikely to be centrally scheduled; conversely, it will be
In practice, what is typically provided is a middleware likely decided based on local interaction of IoT users and de-
platform guaranteeing a pre-dened infrastructure for vices. This in turn may stimulate a game-theoretical ap-
development and execution of distributed applications. proach to the resulting problem of resource (object) sharing
Middleware design, in particular, has become a popular in the IoT.
research area [8991]. Middleware communications may IoT may well inherit concepts and lessons learned in
involve synchronous, asynchronous, message or request- pervasive computing, ambient intelligence applications
oriented methods. The IoT domain spans any of those mod- and service-oriented computing [9699], as detailed in
els depending on the specic application targeted. Section 3. Researchers working in the eld of humancom-
As mentioned before, a key issue of IoT systems will be puter interfaces and user-centric design methodologies, in
the possibility to address objects using unique IDs. The ini- particular, addressed already several issues concerning the
tiative for the denition of a global naming system, ONS impact of sensorized and pervasive environment on the
[92], is meant to extend the concept of Domain Name Ser- user experience [100]. Since IoT will take the reference sce-
vice (DNS) to real-world RFID-tagged objects. Indeed, the narios one step further in terms of scale and offered features,
ability of distinguishing objects is key in enabling distrib- it will also require the development of suitable, scalable ser-
uted applications. At present, the possibility to address an ob- vice delivery platforms that permit multiple services to coex-
ject and a network node in a seamless fashion is a quite deep ist. As mentioned already in the previous sections, in
technical issue and requires a global-scale standardization ef- literature there exists indeed a clear gap as concerns reference
fort, probably wider than ONS. This issue represents also one architecture models able to support the composition of IoT
of the key technical barriers to overcome in order to foster based services.
wide adoption of IoT technology. Another key set of research challenges relate to security
issues. Due to their fundamental role as enablers of IoT
4.3. Distributed intelligence applications, they will be separately discussed in the fol-
lowing section.
Given a system of smart objects that are interconnected The taxonomy of the main research areas and related
in the digital domain and equipped with suitable interfaces topics relevant to the Internet-of-Things, as described
for programming purposes, applications need to coordi- above, is graphically depicted in Fig. 1.
Fig. 1. Taxonomy of research areas relevant to Internet-of-Things.
5. Security denition of an access control mechanism and second,

the denition of an object authentication process (with a
Security represents a critical component for enabling related identity management system).
the widespread adoption of IoT technologies and applica- As data in IoT applications will be related to the physi-
tions. Without guarantees in terms of system-level con- cal realm, ensuring data condentiality is a primary con-
dentiality, authenticity and privacy the relevant straint for many use cases (see Section 6 for further
stakeholders are unlikely to adopt IoT solutions on a large description of potential application scenarios). As a rst
scale. In early-stage IoT deployments (e.g., based on RFIDs example, we may consider data provided by bio-sensors
only), security solutions have mostly been devised in an ad on bacterial composition of the product used for guaran-
hoc way. This comes from the fact that such deployments teeing the required quality in the food industry. This data
were usually vertically integrated, with all components un- is clearly condential because their uncontrolled spreading
der the control of a single administrative entity. In the per- could harm company reputation and its competitive
spective of an open IoT eco-system, whereby different advantage over competing companies. As a second exam-
actors may be involved in a given application scenario ple, we may consider an environmental monitoring appli-
(e.g., one stakeholder owing the physical sensors/actua- cation, whereby data is used to feed an early warning
tors, one stakeholder handling the data and processing system against, e.g., the rise of tsunami/earthquakes, etc.
them, various stakeholders providing different services In such a setting, data should be accessible only by the rel-
based on such data to the end-users, etc.), a number of evant civil protection bodies, which can then put in place
security challenges do arise. In this section, we aim at appropriate risks management strategies. The leakage of
revising and discussing the major security challenges to such information into the public sphere may give rise to
be addressed to turn Internet-of-Things technology into a chaotic and panic situations, putting at risk the safety of
mainstream, widely deployed one. In particular, we identi- large groups of people.
ed three key issues requiring innovative approaches: data Customary solutions for ensuring data condentiality
condentiality, privacy and trust. In the following, we ana- may not be straightforwardly applied to IoT contexts, due
lyze them one by one. It is worth remarking that, as de- to two major limiting factors. The rst one concerns the
picted in Fig. 1, security considerations are orthogonal to sheer amount of data generated by such systems, and re-
the other research areas, and span both the communica- lates hence to scalability issues. The second one relates to
tions/networking, platform/data management and applica- the need of controlling the access to data in an on-line
tion/service levels. and exible way, with access rights changing at run-time
and being applied to dynamic data streams.
Various access control techniques have been proposed
5.1. Data condentiality to ensure condentiality in knowledge management sys-
tems.3 A standard approach, which matches well the fea-
Data condentiality represents a fundamental issue in tures of IoT environments, is represented by Role-Based
IoT scenarios, indicating the guarantee that only autho- Access Control (RBAC) [101]. The concept of RBAC has
rized entities can access and modify data. This is particu- emerged in the past decade as a widely used and highly suc-
larly relevant in the business context, whereby data may cessful alternative to conventional discretionary and manda-
represent an asset to be protected to safeguard competi- tory access controls. In RBAC, users and permissions are
tiveness and market values. In the IoT context not only
users, but also authorized objects may access data. This 3
All techniques are based on a strong trust assumption with respect to
requires addressing two important aspects: rst, the the system platform that handles the access attributes.
assigned to roles. Users acquire permissions indirectly via good starting point for the development of an holistic solu-
roles assignment. The main advantage of RBAC, in an IoT tion for IoT scenarios.
perspective, is the fact that access rights can be modied In many applications aggregated data obtained from
dynamically by changing the role assignments. The IoT con- multiple data sources by applying adequate operators, will
text requires the introduction of new forms of RBAC-style be used. In the literature there are many works that ad-
solutions, in particular considering that IoT data will likely dress security issues of aggregated data in WSN. These
represent streams to be accessed in real-time, rather than works have been classied in hop-by-hop encrypted data
constituting static databases. Data Stream management sys- aggregation and end-to-end encrypted data aggregation.
tems have been increasingly used to support a wide range of In the former the data is encrypted by the sensing nodes
real-time applications (battleeld and network monitoring, and decrypted by the aggregator nodes. The aggregator
sensor networks and so on), and represents a suitable solu- nodes, then, decrypt data coming from the sensing nodes,
tion for the IoT context. In IoT, access control techniques aggregate data and encrypt the aggregated data again. At
should be integrated with data streams management sys- last, the Sink gets the nal encrypted aggregation result
tems. The scientic literature offers few proposals, which and decrypts it. In the end-to-end encrypted data aggrega-
are classied into two main categories: those aiming to en- tion the intermediate aggregator nodes have not the key
sure authenticity, condentiality and integrity of data and can only do aggregations on the encrypted data.
streams during transmission [102,103] and those related to Different hop-by-hop related works [110112] assumes
access control [104,105]. An example of the rst category that data security is guaranteed by means of some key dis-
is presented in [103], which proposes an extension of the tribution schemes. For example SEDAN [113] proposes a
RC4 encryption algorithm to overcome possible decryption secure hop-by-hop data aggregation protocol, in which
failures due to synchronization problems. The proposed each node can verify immediately the integrity of its two
encryption scheme has been developed in the Nile [106] hops neighbors data and the aggregation of the immediate
stream engine. Another interesting proposal is discussed in neighbors by means a management of new type of key,
[102], where authors address the authenticity problem of called two hops pair-wise key. The performance of SEDAN,
outsourced data streams. More precisely, [102] considers a evaluated by means of ad hoc simulation, shows that such
scenario where a data owner constantly outsources its data scheme is able to outperform competitive solutions such as
streams, complemented with additional authentication SAWAN [110] in terms of overhead and mean time to
information, to a service provider. As far as data stream ac- detection. All hop-by-hop proposed solutions are vulnera-
cess control is considered, it is only recently that mecha- ble because the intermediate aggregator nodes are easy to
nisms to guard against unauthorized access to streaming tamper and the sensor readings are decrypted on those
data have been investigated. The work in [104] proposes a aggregators. End-to-end encrypted techniques overcome
model for extending RBAC to protect data streams from this weakness of hop-by-hop techniques. Notice that
unauthorized access. The basic idea is to apply a newly de- end-to-end secure data aggregation techniques also use a
signed operator at the stream, resulting from the evaluation key scheme. Some approaches [114117] suggest to share
of a query to lter out output tuples that do not satisfy ac- a key among all sensing nodes and the Sink, the aggregator
cess control policies. The main drawback of this approach nodes have not the key because the aggregator nodes han-
is that the proposed framework is not able to handle certain dle data without making any encryption/decryption opera-
control policies on views of data from multiple streams, as tion. The limitation of such a solution is that the whole
occurs in IoT. Another relevant work is presented in [105], network is compromised in case the key is compromised
where the authors propose that the data access policies in a sensing nodes.
are dened by the user owning the devices and within the An alternative approach is represented by the adoption
data stream itself. This makes users able to specify how of public-key encryption [118], but in this case the draw-
the data streams management system has to access her/his back is represented by the related high computational cost.
personal data. As such, this solution is more suitable for The aforementioned solutions are all focused on lower
addressing privacy issues, rather than general access control layer security issues, i.e., on the adoption of encryption
problems. This approach is also dependent on the adopted techniques and ad hoc key distribution schemes [119
stream engine, raising issues in terms of support of hetero- 121]. In the IoT domain, the use of aggregated data requires
geneous stream engines. The most general available solu- to address two other fundamental research challenges. The
tion, to best of our knowledge, is [107], which extends the rst one is related to the access control of aggregated data/
work of [108,109] by proposing a general framework to pro- data streams: in case of the aggregation of data with differ-
tect streaming data that is independent form the target en- ent access attributes a solution is needed to establish the
gine. The framework is based on an expressive role-based access attributes of the aggregated data. The second, re-
access control tailored for data streams [108]. It exploits a lated, one deals with the introduction of appropriate oper-
query rewriting mechanism, which rewrites user queries in ators for ensuring the impossibility of recovering raw data
such a way that they do not return data tuples that should streams from the aggregated one.
not be accessed according to the specied access control pol- Furthermore, in order to avoid unauthorized access,
icies. Furthermore, the framework includes a deployment especially considering the use of wireless communications
module that translates the rewritten query in such a way means at the lower layers, the access control mechanisms
that it can be executed by heterogeneous stream engines, should be combined with appropriate data protection
thus overcoming the lack of a standardized stream engine techniques. Typical examples are anonymization tech-
solution. This framework should therefore be considered a niques based on data suppression or randomization
[122,123], or other data cloaking mechanisms, which per- the distributed nature of IdPs and SPs in IoT applications.
turb data following some criteria (e.g., k-anonymity guar- We thus need a secure and privacy-preserving mechanism
antees that every record is indistinguishable from at least for retrieving the entity attributes from different SPs. The
k 1 other records [124]). Relevant issues to be addressed IdM system must provide only the objects information
in this context relate to scalability and energy consump- that is needed to satisfy the requesting SPs access control
tions of existing solutions, which may not meet the policies. In this regard, objects should present different
requirements typical of IoT deployments. accessibility (privacy) levels for various types of informa-
Another aspect that should be considered when the tion. For example, depending on the specic application
problem of condentiality is faced is that of identity man- considered, an object might agree to share a given type
agement. In fact this issue is critical in IoT scenarios where of information, but not all its attributes. Such requirements
there is a fusion of digital and physical world. The problem call for a exible and selective approach to sharing entity
is to nd solutions for handling in secure manner the iden- attributes in federated systems. A system could achieve
tity of objects/things and the related authorization pro- selective release of identity by supporting multiple feder-
cesses. Although users identity management is a well- ated digital identities. In this direction is the proposal of
investigated topic in the literature, managing the identity [125] that integrates federated IdM with trust-negotiation
of smart objects raises a number of novel issues to be dealt techniques. In this way, entities do not have to provide a
with. given attribute more than once to a given federation.
First, it is necessary to reach an agreement on a well- Although it represents a promising approach, for both its
dened concept of identity, when referred to a smart ob- exible/distributed nature and its capability to couple
ject. A well characterized denition of identity should in- identity management with trust, its application to IoT sce-
deed drive the development of an object identity narios require proper tailoring and further studies.
management system (IdM), specifying the main operations Summarizing, the main research challenges for ensur-
that the IdM should perform. Looking at the state-of-the- ing data condentiality in an IoT scenario, as reported in
art, a starting point could be represented by the concept Fig. 2, relate to:
of federation [125]. A federation is dened as a set of orga-
nizations that establish trust relationships with respect to Denition of suitable mechanisms for controlling access
the identity information maintained. A federated identity to data streams generated by IoT devices.
management system provides a group of organizations Denition of an appropriate query language for
that collaborate with mechanisms for managing and gain- enabling applications to retrieve the desired informa-
ing access to identity information of a given entity in the tion out of a data stream.
system and other resources across organizational bound- Denition of a suitable smart objects identity manage-
aries. Traditionally, identity management systems consider ment system.
users as entities whose identity has to be managed; in our
case we are interested in systems whereby the identity 5.2. Privacy
attributes relate to smart objects, and not to users.
IdM systems involve at least two types of actors: iden- Privacy denes the rules under which data referring to
tity providers (IdP) and service providers (SP). An IdP man- individual users may be accessed. The main reasons that
ages authentication of entities4 and of entity-relevant makes privacy a fundamental IoT requirement lies in the
information. A SP offers services to users that satisfy the pol- envisioned IoT application domains and in the technolo-
icy requirements associated with the offered services. It fur- gies used. Health-care applications represent the most out-
ther species and enforces the access control policies for the standing application eld, whereby the lack of appropriate
resources it offers. An organization in a federation can act as mechanisms for ensuring privacy of personal and/or sensi-
both an IdP and a SP. tive information has harnessed the adoption of IoT tech-
In most IdM systems, IdPs authenticate entities using nologies. In addition, in the IoT vision, a prominent role
single-sign-on (SSO) technology. With SSO, conventionally, will be played by wireless communication technologies.
users can log on with the same user name and password The ubiquitous adoption of the wireless medium for
for seamless access to federated services within one or exchanging data may pose new issue in term of privacy
multiple organizations. Federated identity includes not violation. In fact, wireless channel increases the risk of vio-
only users login names, but also user properties, or user lation due to the remote access capabilities, which poten-
identity attributes (user attributes, for short). Thus, autho- tially expose the system to eavesdropping and masking
rizations, specied for a given resource, are no longer ex- attacks. Hence privacy represents a real open issue that
pressed in terms of user login IDs, but in terms of may limit the development of the IoT.
requirements and conditions against user properties. A number of frameworks have been proposed for
In order to apply these concepts to IoT scenarios, we accounting for privacy issues in the system design phase,
need to assess their suitability to deal with smart objects such as Kaos [126], Tropos [127,128], NFR [129,130],
instead of users. Further, we need to properly account for GBRAM [131], PRIS [132,133]. The latter approach may
represent a viable starting point for the denition of appro-
4
priate privacy-preserving mechanisms for IoT. PRIS [132],
Traditional IdM systems handle identities of users. As we are interested
also in handling identities of smart objects, we use the term entity in the
indeed, represents a requirement engineering methodol-
remainder to indicate both users and smart objects, depending on the ogy, which incorporates privacy requirements into the sys-
application context. tem design process. PRIS provides a set of concepts to
Data stream access control,

Identity management
Data system, Confidentiality-
Confidentiality preserving aggregation
General privacy model for

IoT, Enforcement
Privacy mechanisms, Role-based
systems, Data governance,
Trust negotiation
Trust mechanisms, Negotiation
language, Object identity
Fig. 2. Graphical representation of security challenges in Internet-of-Things.
model privacy requirements and a set of rules to transform 5.3. Trust

such requirements into implementation techniques. Differ-
ent is the goal of [133], which denes a general UML con- The concept of trust is used in a large number of differ-
ceptual model for representing privacy policies. The model ent contexts and with diverse meanings. Trust is a complex
species the needed functional modules of an application notion about which no consensus exists in the computer
in order to enforce such policies, introducing all the ele- and information science literature, although its importance
ments required for the denition of privacy aware systems. has been widely recognized. Different denitions are possi-
As it operates at a very high level of abstraction, it is suit- ble depending on the adopted perspective. A main problem
able for application to IoT scenarios, characterized by a with many approaches towards trust denition is that they
high degree of heterogeneity in terms of privacy do not lend themselves to the establishment of metrics and
requirements. evaluation methodologies.
At the same time, the development of concrete ap- A widely used denition is the one provided by Blaze
proaches for building privacy-preserving mechanisms for and Feigenbaum [134], which refers to security policies
IoT applications still presents a number of challenging as- regulating accesses to resources and credentials that are
pects. The development of concrete implementations required to satisfy such policies. Trust negotiation refers
would benet from the denition of a general model, able to the process of credential exchanges that allows a party
to represent all IoT fundamental entities and their relation- requiring a service or a resource from another party to pro-
ships. Moreover the implementations should include vide the necessary credentials in order to obtain the ser-
enforcement mechanisms able to cope with the scale and vice or the resource. This denition of trust is very
with the dynamic nature of IoT scenarios. In order to sat- natural for secure knowledge management as systems
isfy such requirements, solutions also able to enforce a dy- may have to exchange credentials before sharing knowl-
namic data stream access control should be provided. edge. For this reason, we base our analysis of trust issues
Summarizing, the open research challenges in terms of pri- in IoT upon it. Trust negotiation relies on peer-to-peer
vacy-preserving mechanisms for IoT, as reported in Fig. 2, interactions, and consists of the iterative disclosure of dig-
are given by: ital credentials, representing statements certied by given
entities, for verifying properties of their holders in order to
Denition of a general model for privacy in IoT. establish mutual trust. In such an approach, access re-
Development of innovative enforcement techniques, sources (data and/or services) is possible only after a suc-
able to support the scale and heterogeneity characteriz- cessful trust negotiation has been completed. A trust
ing IoT scenarios. negotiation system typically exploits digital identity infor-
Development of solutions that balance the need of ano- mation for the purpose of providing a ne-grained access
nymity presented by some applications with the locali- control to protected resources. The ability to meet the trust
zation and tracking requirements of some other ones. requirement is indeed strictly related to the identity man-
This entails the denition of privacy policies, that spec- agement and access control issues, as discussed above. At
ify under which conditions it is possible to identify and present a limited number of solutions are available
localize a smart object. Moreover, it needs to specify [135,136,46,134]. The most popular approaches include
when it is possible to access sensitive data. KeyNote [134] and TrustBuilder [46], which nonetheless
do not lend themselves to a straightforward application to the adoption of RFID technology by major governments
the IoT domain, due to the high computational require- (e.g., the issuance of RFID-tag-inlaid Resident ID cards by
ments they impose. Many open issues have to be ad- the Ministry of Public Security of China).
dressed in order to develop IoT trust services. First, the The increase in the usage of RFID, paving the way to
denition of globally accepted certication authorities making Internet-of-Things a reality, is not simply a result
should be addressed, together with a number of require- of technological push; it is also driven by the market pull,
ments that an IoT-compliant certication authority should since enterprises are increasingly realizing the commercial
respect. Furthermore, it is necessary to devise an effective benets of applications that can be realized with Internet-
trust negotiation language, able to simplify credential of-Things technologies. The evolution of Internet-of-Things
specications and to express a wide range of protection may follow the evolution path of mobile phones [137]. At
requirements through the denition of exible disclosure the end of 2009 there were 5.9 billion mobile phone sub-
policies. In addition, the denition of an effective model scribers (source: ITU), driven by the need to communicate
of trust should account for both the highly distributed nat- anywhere and at anytime. Now, imagine this connectivity
ure of the IoT as well as for the requirements (in terms of being brought to everyday objects: fridges, cars, cups, keys,
computational complexity and/or response time) typical etc., as it will be enabled by IoT. A huge market opportunity
of many IoT applications. exists for Internet-of-Things, related to the possibility of
In other words, we need to move away from the classi- networking smart things and of providing applications
cal centralized and static approaches underpinning the leveraging said connectivity.
most widely used trust management solutions, to adopt a Besides enhancing the competitiveness of various verti-
fully distributed and dynamic approach that assumes that cal markets, IoT technologies can open up new business
no trust relationship is dened a priori among the entities opportunities by: (i) bridging vertical markets, giving rise
in the system. Moreover, a new exible framework for to cross-cutting applications and services, based on the
trust management should be introduced in order to meet use of a common underlying ICT platform, (ii) enabling
the scalability requirements that arise at different levels, the arising and growth of new market segments and appli-
including, e.g., naming and addressing information knowl- cations, made possible by the ability, provided by IoT tech-
edge management and service provisioning. nologies, to interact with physical objects via digital means
Anyway, although the complete dynamic and distrib- and (iii) optimizing business processes by leveraging on
uted nature of IoT makes to address trustworthiness extre- advanced analytics techniques applied to IoT data streams.
mely challenging, we may well consider IoT as an As an example of the latter point, we could consider the
extremely interesting application of trust concepts. In fact smart fridge scenario, whereby items stored in a refriger-
in a context in which smart objects themselves take deci- ator are identied by means of RFID or equivalent technol-
sions, the rst trust relationship has to be established ogies and the fridge has embedded computing and
among humans and the objects surrounding them. networking capability, so that it may understand the quan-
The most relevant research challenges in the denition tity and type of items stored and decide whether there is a
of appropriate trust mechanisms for IoT, as reported in need to buy new items, etc. At the moment the electronic
Fig. 2, can be summarized as: appliances and the large-scale retail trade represent sepa-
rate industrial sectors. Without a set of common technical
Introduction of a simple trust negotiation language sup- standards and interfaces (at both the device and semantic
porting the semantic interoperability requirements of level) joining the activities of such two sectors, an IoT-en-
IoT. abled device like the smart fridge could not take place. Vice
Denition of a trust negotiation mechanism based on a versa, the adoption of IoT technologies can give rise to new
ne-grained access control of data streams. business ecosystems, characterized by new actors and
Development of an adequate object identity manage- value chains. An example could be a brokerage service that,
ment system. by accounting for what is currently in your fridge, your die-
Design of a general and exible trust management tary constraints and tastes, your agenda (in terms of din-
framework able to leverage the aforementioned items. ners with friends, etc.) negotiates for you the best food at
the best rate, etc.
6. Applications and impact areas In terms of application elds and market sectors where
IoT solutions can provide competitive advantages over cur-
The concept of Internet-of-Things, with its vision of rent solutions, we identied six ones which we do believe
Internet-connected objects of various capabilities and form can play a leading role in the adoption of IoT technologies:
factors, could boost the role of ICT as innovation enabler in environmental monitoring; smart cities; smart business/
a variety of application markets. inventory and product management; smart homes/smart
One of the technological pillars of the Internet-of-Things, building management; health-care and security and sur-
namely RFID technology, has already been incorporated into veillance. In the following we briey discuss the relevance
a wide array of products. The number of RFID tags sold in and potential impact of IoT technologies on the competi-
2011 accounted to 2.88 tags (source: www.idtechex.com), tiveness of players in such markets.
with an estimated market value of $ 5.84 billions (source:
www.idtechex.com). Adoption of RFID technology in indus- Smart Homes/Smart Buildings. Instrumenting buildings
try slowed down in 2008/2010 as a consequence of the glo- with advanced IoT technologies may help in both
bal economic downturn, but this decrease got balanced by reducing the consumption of resources associated to
buildings (electricity, water) as well as in improving the enable access to critical areas, whereby the presence
satisfaction level of humans populating it, be it workers of human operators might not represent a viable option
for ofce buildings or tenants for private houses. Impact (e.g., volcanic areas, oceanic abysses, remote areas),
is both in economic terms (reduced operational expen- from where sensed information can be communicated
ditures) as well as societal ones (reducing the carbon to a decision point in order to detect anomalous condi-
footprint associated to buildings, which are a key con- tions. In this perspective, IoT technologies can enable
tributors to the global greenhouse gas emissions). In the development of a new generation of monitoring
this application, a key role is played by sensors, which and decision support systems, providing enhanced
are used to both monitor resource consumptions as well granularity and real-time capabilities over current solu-
as to proactively detect current users needs. Such a sce- tions. Another case in which the sensing ability of IoT
nario integrates a number of different subsystems, and devices supports the environmental safety is repre-
hence requires a high level of standardisation to ensure sented by re detection. When a suite of sensors detects
interoperability. Ability to reason in a distributed, coop- the possible presence of re (by means, e.g., of temper-
erative way, and to actuate is also necessary in order to ature sensors), an alarm is sent directly to the re
ensure that decisions taken on the resources under con- department in a short time (exploiting the advanced
trol (e.g., switch on/off lighting, heating, cooling, etc.) communication features of IoT platform), along with
are in line with the users needs and expectations, other parameters that are useful in decision making
which in turn are strictly intertwined to the activities and support, such as the description of the area subject
they undertake and/or plan to take. to the re, the possible presence of people, of inamma-
Smart Cities. The term Smart Cities is used to denote ble materials, etc. Clearly, rapid response has the conse-
the cyberphisycal eco-system emerging by deploying quence of saving human lives, mitigating the damage to
advanced communication infrastructure and novel ser- the property or vegetation and in general reducing the
vices over city-wide scenarios. By means of advanced level of disaster. Many other scenarios related to civil
services, it is indeed possible to optimize the usage of protection can prot from IoT technologies (tunnel area,
physical city infrastructures (e.g., road networks, power earthquake, tsunami, etc.), whereby the ability to access
grid, etc.) and quality of life for its citizens. IoT technol- environmental data in real-time over large-scale areas
ogies can nd a number of diverse application in smart enable the uptake of efcient coordination strategies
cities scenarios. As a case study, IoT technologies can be among rescue teams.
used to provide advanced trafc control systems. Health-care. IoT technologies can nd a number of appli-
Through IoT it will be possible to monitor car trafc in cations in the health-care sector. On the one hand, they
big cities or highways and deploy services that offer can be used to enhance current assisted living solutions.
trafc routing advice to avoid congestion. In this per- Patients will carry medical sensors to monitor parame-
spective, cars will be understood as representing smart ters such as body temperature, blood pressure, breathing
objects. In addition, smart parking devices system, activity. Other sensors, either wearable (e.g., accelerom-
based on RFID and sensor technologies, may allow to eters, gyroscopes) or xed (proximity) will be used to
monitor available parking spaces and provide drivers gather data used to monitor patient activities in their liv-
with automated parking advice, thus improving mobil- ing environments. Information will be locally aggregated
ity in urban area. Moreover, sensors may monitor the and transmitted to remote medical centers, which will be
ow of vehicular trafc on highways and retrieve aggre- able to perform advanced remote monitoring and will be
gate information such as average speed and numbers of capable of rapid response actions when needed. The
cars. Sensors could detect the pollution level of air, interconnection of such heterogeneous sensors could
retrieving smog information such as the level of carbon provide a comprehensive picture of health parameters,
dioxide, PM10, etc., and deliver such information to thereby triggering an intervention by the medical staff
health agencies. Furthermore, sensors could be used in upon detection of conditions that may lead to health
a forensics setting, by detecting violations and by trans- deterioration, thus realizing preventive care.
mitting the relevant data to law enforcement agencies Another relevant application sector relates to personal-
in order to identify the violator, or to store information ized health-care and well-being solutions. The use of
that will be provided in case of accident for subsequent wearable sensors, together with suitable applications
accident scene analysis. running on personal computing devices enables people
Environmental monitoring. IoT technology can be suit- to track their daily activities (steps walked, calories
ably applied to environmental monitoring applications. burned, exercises performed, etc.), providing sugges-
In this case a key role is played by the ability of sensing, tions for enhancing their lifestyle and prevent the onset
in a distributed and self-managing fashion, natural phe- of health problems.
nomena and processes (e.g., temperature, wind, rainfall, Smart business/Inventory and product management. RFID
river height), as well as to seamlessly integrate such technologies are already used in many sectors for
heterogeneous data into global applications. Real-time inventory management, throughout the supply and
information processing, coupled with the ability of a delivery chain. This relies on the ability of RFID technol-
large number of devices to communicate among them, ogies to identify and provide support for tracking goods.
provides a solid platform to detect and monitor anoma- At the moment, however, RFID applications are built in
lies that can lead to endangering human and animal life. a rather ad hoc fashion, and are only partially integrated
The vast deployment of miniaturized devices may into supply management systems.
RFID are customarily used to monitor and manage the improve the competitiveness of the solutions at hand. IoT
movement of products through a supply chain; typi- adoption is therefore expected to be to strongly driven by
cally, RFID tags are directly attached to the items (or the market needs and by the market dynamics. At the same
to the containers that carry them), while readers are time, ICT industries, standardisation bodies and policy-
placed throughout the facility to be monitored. IoT makers are undertaking a series of initiatives to steer the
technologies can provide enhanced exibility in terms IoT development process with the objective of maximizing
of readers positions, while at the same time enabling its socio-economic value while minimizing the threats re-
seamless interoperability between RFID-based applica- lated to privacy and condentiality of data. In this regard,
tions used by different actors dealing with the product the following section reviews a number of IoT initiatives
throughout the various phases of its life-cycle. and also provides a discussion on standardisation
In retail applications, IoT technologies can be used to activities.
monitor in real-time product availability and maintain
accurate stock inventory. They can also play a role in
7. Related on-going initiatives
after-market support, whereby users can automatically
retrieve all data about the products they bought. Also,
A number of large-scale initiatives on IoT are active in
identication technologies can help in limiting thefts
the US, in Europe, in Japan, China, Korea and other coun-
and in ghting counterfeiting by providing products
tries. In the following subsection we will briey report on
with a unique identier including a complete and trust-
the most relevant ones.
worthy description of the good itself.
Besides research initiatives, standardization activities
Furthermore, sensors and specically bio-sensor tech-
are also of key importance in order to ensure a successful
nologies in combination with RFID technology may
widespread adoption of IoT technologies and services. In
allow control production processes, nal product qual-
Section 7.2 we briey report on the most relevant ones.
ity and possible shelf life deterioration of the product,
e.g., in the food industry. For example, RFID devices
can be used to identify and track the product, while 7.1. IoT related projects
the bio-sensors can monitor parameters such as tem-
perature and bacterial composition in order to guaran- The growing interest in IoT technologies and applica-
tee required quality of the nal product. tions is well exemplied by the number of research initia-
Security and surveillance. Security surveillance has tives arising worldwide around such themes. In the US, the
become a necessity for enterprise buildings, shopping American National Science Foundation (NSF) launched in
malls, factory oors, car parks and many other public 2008 a program on Cyber-Physical Systems,5 aimed at
places. Homeland security scenarios faces also similar introducing systems able to merge computational and phys-
threats, albeit on a different scale. IoT-enabled technol- ical resources. The program is meant to cover a wide array of
ogies can be used to greatly enhance the performance of application scenarios, ranging from smart electric grid to
current solutions, providing cheaper and less invasive smart transportation, from smart medical technologies to
alternatives to the widespread deployment of cameras smart manufacturing. The 2010 report of the Presidents
while at the same time preserving users privacy. Ambi- Council of Advisors on Science and Technology, Designing
ent sensors can be used to monitor the presence of dan- a digital future: federally funded research and development
gerous chemicals. Sensors monitoring the behaviour of in networking and information technology6 encourages fur-
people may be used to assess the presence of people ther investments in Cyber-Physical System, due to their high
acting in a suspicious way. Efcient early warning sys- potential impact on a number of critical industrial sectors.
tems can therefore be built. Personal identication by The European Commission has been pushing initiatives
means of RFID or similar technologies is also an option. related to IoT since 2005 [5], and has recently launched, in
However in many countries user associations are er- the framework of the 7th Framework Programme, an ini-
cely protesting about the privacy infringement that tiative on Internet-Connected Objects. The focus is on
could result from the widespread adoption of such a adoption of IoT technologies and services in enterprise
technology. When used in conjunction with role-based environments, with the aim of increasing the competitive-
access control systems, IoT technologies can provide ness of European industry through adoption of IoT-enabled
high level of exibility, being able to cope with access solutions7 http://www.rd-in-action.eu/cerp. Activities in
policies (e.g., to different areas of buildings) which such eld led to the denition of a strategic research agenda,
may change over time due to logistic changes and/or including a description of European strategies in this sector
to changes in role of the user and/or according to con- [6].
textual information (e.g., some areas not accessible on Within the initiatives which have taken place at the
a given day due to renovation works going on). Also in European level, four large-scale ones are worth mention-
this market the advantages are in terms of enhanced ing. The HYDRA project8 developed a middleware based
functionality, better user acceptance through reduction on a service-oriented architecture, transparent to the under-
of the use of cameras, reduced operational costs and 5
http://www.nsf.gov/funding/pgm_summ.jsp?pims_id=503286.
increased exibility in a changing environment. 6
http://www.whitehouse.gov/sites/default/les/microsites/ostp/pcast-
nitrd-report-2010.pdf.
Clearly, the scope of IoT is extremely wide. However, 7
http://cordis.europa.eu/fp7/ict/programme/challenge1_en.html.
applications that are built on top of IoT may consistently 8
http://www.hydramiddleware.eu.
lying communication, supporting distributed as well as cen- der to enable pervasive services, with target applications
tralized architectures, security and trust models. This project ranging from smart home environments to supply chain
was meant to provide a middleware solution allowing the management.
developers to incorporate heterogeneous physical devices While the widespread diffusion of research initiatives
into their applications by offering easy to use Web service denotes the vitality of the eld and the potential of IoT
interfaces for controlling the physical devices. Support was applications, it brings alongside a risk of fragmentation
provided for a number of underlying communication tech- and of lack of adoption of adequate standards. IoT would
nologies, including Bluetooth, RF, ZigBee, RFID, WiFi, etc. require, as the technology gets mature and makes its way
The Hydra middleware included methods for performing into the real world, a careful standardization process, in or-
effectively device and service discovery, for supporting der to ensure interoperability among devices and applica-
peer-to-peer interaction models and efcient diagnostics tions coming from different countries, building the
tools. Solutions for distributed security and social trusts foundations the real arising of an Internet of things.
were also devised and prototyped.
The RUNES project9 was meant to create a large-scale, 7.2. Standardization activities
widely distributed, heterogeneous networked embedded
systems which provide a exible and adaptable ICT tool to A number of standardization activities with focus on
leverage environmental data. The main target of RUNES is tag-based technologies has been active in the last years.
a fully operational middleware enabling the potential for These standardization activities are conned mostly within
the introduction of a new class of networked embedded sys- the sensing/RFID domain. In particular, the RF-layer and
tems. In RUNES, one of the target challenges was to achieve the NFCIP (Near Field Communication Interface and Proto-
the required level of self-organization to suit a dynamic col) are already standardized by various bodies (ISO 18092,
environment, while ensuring that proper interfaces were 21481, 22536 and 23917; ECMA 340, 352, 356 and 365;
provided to programmers in order to ease the development ETSI TS 102 190). Also, ECMA 340/352 and ISO 18092/
of applications and services. This was meant to allow for a 21481 describe the Near Field Communication Interface
signicant cut in the cost of new application development and Protocol (NFCIP-1 and -2). Test methods for interfaces
and a much faster time to market. and protocols are described in ECMA 356/362 and in ISO
The IoT-A project10 aims at introducing an architectural 22536/23917. In parallel, also the Global System for Mobile
reference model for the interoperability of Internet-of- Communications Association (GMSA) established a NFC
Things, together with a set of mechanisms for its efcient working group in 2006 and already derived guidelines for
integration into the service layer of the Future Internet. NFC services to be supported by cellular phones technolo-
The project is a large-scale one, involving a number of rele- gies. The reason for the interest of the GSMA is that cellular
vant stakeholders and addressing a number of application technology is perceived as a potential enabler for the diffu-
domains. Particular attention is paid to resolution schemes, sion of a large number of services based on the use of
whereby innovative approaches are proposed to ensure scal- embedded NFC devices (e.g., micro-payments).
able look-up and discovery of smart objects and associated A key issue in IoT relate to the naming systems. As most
resources. IoT applications would require unique identiers, a global
The iCORE project11 aims at empowering the IoT with coordination of the naming scheme to be used is needed.
cognitive technologies and is focused around the concept In the RFID eld, the most widely adopted solution is the
of virtual objects (VOs), intended as semantically enriched Electronic Product Code (EPC). Specication of EPC identi-
virtual representation of the capabilities/resources provided ers constitute an open, freely accessible standard, issued
by real-world objects fostering their re-usability and sup- by EPCglobal Inc., and based upon the work carried out
porting their aggregation into more composite services in the last decade at the MIT Auto-ID Center. Object Nam-
(composite virtual objects CVOs). VOs provide a unied ing Service (ONS) represents a mechanism for discovering
representation, thereby hiding any underlying technological information about a given object starting from its EPC [92].
heterogeneity and providing a standardized way of access- In terms of communications among smart objects, we
ing objects capabilities and resources. One key element in should distinguish between two aspects. At the lower lay-
the iCORE project is the use of advanced cognitive tech- ers (PHY and MAC), IEEE is running the 802.15 Working
niques for managing and composing VOs to improve IoT Group on wireless personal area networks. This includes
applications and better match user/stakeholder require- a number of task groups, which led to the denition of
ments. Four use cases are put forward for validation pur- the 802.15.4 specications, which are at the basis of the
poses: ambient assisted living, smart ofce, smart ZigBee technology. Lately, attention has also been devoted
transportation and supply chain management. to optical wireless communications, within the 802.15.7
IoT-centric programs are active also in Japan, under the Task Group. In terms of upper layers, ETSI has launched
umbrella of the UNS initiative (Ubiquitous Networked in 2008 a technical committee on Machine-to-Machine
Society, part of the wider e-Japan strategy12), which fo- (M2M) communications, which is however mostly focused
cuses on the ubiquitous presence of sensors and RFIDs in or- on the telecommunications perspective.13
It is important to remark that there is a clear lack of
9 standardization activities related to the data models, ontol-
http://www.ist-runes.org/.
10
http://www.iot-a.eu/. ogies and data format (s) to be used in IoT applications and
11
http://www.iot-icore.eu/.
12 13
http://www.kantei.go.jp/foreign/it/network/0122full_e.html. http://www.etsi.org/Website/Technologies/M2M.aspx.
in terms of service-level interfaces and protocols. Such is- [8] G. Roussos, V. Kostakos, RFID in pervasive computing: state-of-the-
art and outlook, Pervasive Mob. Comput. 5 (2009) 110131. http://
sues are expected to play a key role for enabling semantic
dx.doi.org/10.1016/j.pmcj.2008.11.004.
interoperability and thus the mushrooming of IoT-based [9] I. Akyilidiz, W. Su, Y. Sankarasubramaniam, E. Cayirci, Wireless
services and applications. A working group on Semantic sensor network: a survey, Comput. Netw. 38 (4) (2002) 393422.
Sensor Network was active at W3C from 2009 to 2011; [10] I.F. Akyildiz, I.H. Kasimoglu, Wireless sensor and actor networks:
research challenges, Ad Hoc Netw. J. 2 (2004) 351367.
the nal report could be found here: http://www.w3.org/ [11] J. Saltzer, D. Reed, D. Clark, End-to-end arguments in system design,
2005/Incubator/ssn/XGR-ssn-20110628/. The lack of a ACM Trans. Comput. Syst. 2 (1984) 277288.
shared approach towards such issues could represent a [12] V. Jacobson, D.K. Smetters, J.D. Thornton, M.F. Plasee, N. Briggs, R.
Braynard, Networking named content, in: Proceedings of ACM
barrier for the development of an open IoT architecture, CoNEXT, Rome, Italy, 2009, pp. 112.
thereby harnessing the disruptive innovation potential of [13] S. Dobson, S.G. Denazis, A. Fernndez, D. Gati, E. Gelenbe, F.
such technology. Massacci, P. Nixon, F. Saffre, N. Schmidt, F. Zambonelli, A survey of
autonomic communications, TAAS 1 (2) (2006) 223259. http://
doi.acm.org/10.1145/1186778.1186782.
8. Conclusions [14] W. Elmenreich, R. DSouza, C. Bettstetter, H. de Meer, A survey of
models and design methods for self-organizing networked systems,
in: IWSOS, 2009, pp. 3749.
The Internet-of-Things may represent the next big leap [15] D. Guinard, V. Trifa, F. Mattern, E. Wilde, From the Internet of
ahead in the ICT sector. The possibility of seamlessly merg- Things to the Web of Things: Resource Oriented Architecture and
Best Practices, Springer, New York, Dordrecht, Heidelberg, London,
ing the real and the virtual world, through the massive 2011 (Chapter 5).
deployment of embedded devices, opens up new exciting [16] D. Guinard, V. Trifa, S. Karnouskos, P. Spiess, D. Savio, Interacting
directions for both research and business. with the SOA-based Internet of Things: discovery, query, selection,
and on-demand provisioning of Web services, IEEE Trans. Serv.
In this survey article, we provided an overview of the Comput. 3 (3) (2010) 223235.
key issues related to the development of IoT technologies [17] L. Chen, M. Tseng, X. Lian, Development of foundation models for
and services. A number of research challenges has been Internet of Things, Front. Comput. Sci. China 4 (2010) 376385.
[18] F. Michahelles, F. Thiesse, A. Schmidt, J.R. Williams, Pervasive RFID
identied, which are expected to become major research and near eld communication technology, IEEE Pervasive Comput.
trends in the next years. The most relevant application 6 (3) (2007) 9496.
elds have been presented, and a number of use cases [19] C. Ghezzi, F. Pacici, Evolution of software composition
mechanisms: a survey, in: D. Lucia, F. Ferrucci, G. Tortora, M.
identied.
Tucci (Eds.), Emerging Methods, Technologies, and Process
We do hope that this survey will be useful for research- Management in Software Engineering, J. Wiley and Sons, New
ers and practitioners in the eld, helping them to under- York, 2008, pp. 319.
stand the huge potential of IoT and what are the major [20] S. Haykin, Cognitive radio: brain-empowered wireless
communications, IEEE J. Sel. Areas Commun. 23 (2005) 201220.
issues to be tackled, devising innovative technical solu- [21] I. Chlamtac, M. Conti, J.J.-N. Liu, Mobile ad hoc networking:
tions able to turn IoT from a research vision into reality. imperatives and challenges, Ad Hoc Netw. 1 (1) (2003) 1364.
[22] M. Murphy, J. Butler, Proactive computing: RFID & sensor networks,
Final Report on the Conference organised by DG Information
Acknowledgments Society and Media, Networks and Communication Technologies
Directorate, March 2006. <ftp://ftp.cordis.europa.eu/pub/ist/docs/
ka4/au_conf670306_murphy_en.pdf>.
The work of D. Miorandi and F. De Pellegrini was sup- [23] M. Durvy, J. Abeill, P. Wetterwald, C. OFlynn, B. Leverett, E.
ported by the EC within the framework of the BIONETS Gnoske, M. Vidales, G. Mulligan, N. Tsiftes, N. Finne, A. Dunkels,
Project IST-FET-SAC-FP6-027748, www.bionets.eu. The Making sensor networks ipv6 ready, in: Proceedings of the Sixth
ACM Conference on Networked Embedded Sensor Systems (ACM
authors would like to acknowledge Dr. V. Osmani and Dr.
SenSys 2008), poster session, Raleigh, North Carolina, USA, 2008,
G. Russello for the comments provided while preparing pp. 421422.
this work. The authors also acknowledge the iCore Consor- [24] V. Raghunathan, S. Ganeriwal, M. Srivastava, Emerging techniques
for long lived wireless sensor networks, IEEE Commun. Mag. 44
tium for the feedback and suggestions provided on an early
(2006) 108114.
version of the article. [25] O.B. Akan, M.T. Isik, B. Baykal, Wireless passive sensor networks,
IEEE Commun. Mag. 47 (2009) 9299.
[26] I. Akyilidiz, F. Brunetti, C. Blazquez, Nanonetworking: a new
References communication paradigm, Comput. Netw. 52 (12) (2008) 2260
2279.
[1] M. Weiser, The computer for the 21st century, Sci. Am. (1991) 94 [27] H. Liu, M. Bolic, A. Nayak, I. Stojmenovic, Taxonomy and challenges
100. of the integration of RFID and wireless sensor networks, IEEE Netw.
[2] L. Atzori, A. Iera, G. Morabito, The Internet of Things: a survey, 22 (2008) 2635.
Comput. Netw. 54 (15) (2010) 27872805. [28] L. Zhang, Z. Wang, Integration of RFID into wireless sensor
[3] The Internet of Things, ITU Internet Reports, 2005. <http:// networks: architectures, opportunities and challenging problems,
www.itu.int/internetofthings/>. in: Proceedings of GCCW, 2006, pp. 463469.
[4] M. Zorzi, A. Gluhak, S. Lange, A. Bassi, From todays INTRAnet of [29] E. Aarts, R. Wichert, Ambient intelligence, in: H.-J. Bullinger (Ed.),
things to a future INTERnet of things: a wireless- and mobility- Technology Guide, Springer, Berlin, Heidelberg, 2009, pp. 244
related view, IEEE Wireless Commun. 17 (6) (2010) 4451 . 249.
[5] J. Buckley, From RFID to the Internet of things: pervasive networked [30] J. Gosling, B. Joy, G. Steele, The Java Language Specication,
systems, Final Report on the Conference organised by DG Addison-Wesley, 1996.
Information Society and Media, Networks and Communication [31] R. Grimm, T. Anderson, B. Bershad, D. Wetherall, A system
Technologies Directorate, March 2006. <ftp://ftp.cordis.europa.eu/ architecture for pervasive computing, in: Proceedings of the 9th
pub/ist/docs/ka4/au_conf670306_buckley_en.pdf>. ACM SIGOPS European Workshop, Kolding, Denmark, 2000, pp.
[6] Internet of Things: Strategic Research Agenda, September 2009. 177182.
<http://ec.europa.eu/information_society/policy/rd/documents/ [32] J.P. Sousa, D. Garlan, Aura: an architectural framework for user
in_cerp.pdf>. mobility in ubiquitous computing environments, in: Proceedings of
[7] G. Kortuem, F. Kawsar, V. Sundramoorthy, D. Fitton, Smart objects the 3rd Working IEEE/IFIP Conference on Software Architecture,
as building blocks for the internet of things, IEEE Internet Comput. Kluwer Academic Publishers, Madison, Wisconsin, 2002, pp. 2943.
14 (2010) 4451.
[33] G. Judd, P. Steenkiste, Providing contextual information to [58] D.S. Scherber, H. Papadopoulos, Distributed computing of averages
pervasive computing applications, in: Proceedings of the IEEE over ad-hoc networks, IEEE JSAC 23 (4) (2005) 755764.
International Conference on Pervasive Computing (PERCOM03), [59] S. Barbarossa, G. Scutari, A. Swami, Distributed detection and
Dallas, 2003, pp. 2943. estimation in decentralized sensor networks: and overview, in:
[34] B. Schilit, M. Theime, Disseminating active map information to Proceedings of EURASIP EUSIPCO, Florence, 2006.
mobile hosts, IEEE Netw. 8 (1994) 2232. [60] M.Z.H.Y.E. Bottega, P. Popowsky, R. Prasad, Hypothesis testing over
[35] B. Schilit, N. Adams, R. Want, Context-aware computing a random access channel in wireless sensor networks, in:
applications, in: Proceedings of the IEEE Workshop on Mobile Proceedings of EURASIP EUSIPCO, Florence, Italy, 2006.
Computing Systems and Applications, Santa Cruz, CA, 1994, pp. 85 [61] G. Mergen, L. Tong, Type based estimation over multiaccess
90. channels, IEEE Trans. Signal Process. 54 (2) (2006) 613626.
[36] B. Brumitt, S. Shafer, Topological world modeling using semantic [62] R. Gallagher, Finding parity in a simple broadcast network, IEEE
spaces, in: Proceedings of Workshop on Location Modeling for Trans. Inform. Theory 34 (2) (1988) 176179.
Ubiquitous Computing, Atlanta, GA, 2001, pp. 5562. [63] E. Kushilevitz, Y. Mansour, Computation in noisy radio networks,
[37] M. Baldauf, S. Dustdar, F. Rosenberg, A survey on context-aware in: Proceedings of ACM-SIAM Symposium on Discrete Algorithms,
systems, Int. J. Ad Hoc Ubiquitous Comput. 2 (4) (2007) 263277. San Francisco, USA, 1998, pp. 236243.
[38] P. Oreizy, M. Gorlick, R. Taylor, D. Heimbigner, G. Johnson, N. [64] A. Giridhar, P.R. Kumar, Towards a theory of in-network
Medvidovic, A. Quilici, D. Rosenblum, A. Wolf, An architecture- computation in wireless sensor networks, IEEE Commun. Mag. 44
based approach to self-adaptive software, IEEE Intell. Syst. 14 (4) (2006) 98107.
(1999) 5462. [65] B. Warneke, M. Last, B. Liebowitz, K.S.J. Pister, Smart dust:
[39] J. Coutaz, C. J, S. Dobson, D. Garlan, Context is key, Commun. ACM communicating with a cubic-millimeter computer, Computer 34
48 (2005) 4953. (1) (2001) 4451.
[40] A. Krause, A. Smailagic, D.P. Siewiorek, Context-aware mobile [66] A. Harter, A. Hopper, A distributed location system for the active
computing: learning context dependent personal preferences ofce, IEEE Netw. 8 (1) (1994) 6270.
from wearable sensor array, IEEE Trans. Mob. Comput. 5 (2006) [67] G.D. Abowd, C.G. Atkeson, J. Hong, S. Long, R. Kooper, M. Pinkerton,
113127. Cyberguide: a mobile context-aware tour guide, Wireless Netw. 3
[41] A.K. Clear, R. Shannon, T. Holland, A. Quigley, S. Dobson, P.Nixon, (1997) 421433.
SITUVIS: a visual tool for modelling a users behavior patterns in a [68] P. Vicaire, T. He, T. Yan, Q. Cao, G. Zhou, L. Gu, L. Luo, R. Stoleru, J.A.
pervasive environment, in: Proceedings of the Seventh Stankovic, T. Abdelzaher, Achieving long-term surveillance in
International Conference on Pervasive Computing, Nara, Japan, vigilne, ACM Trans. Sens. Netw. 5 (2009) 139.
2009, pp. 327341. [69] D. Moore, J. Leonard, D. Rus, S. Teller, Robust distributed network
[42] S. Loke, Representing and reasoning with situations for context- localization with noisy range measurements, in: Proceedings of
aware pervasive computing:a logic programming perspective, The SenSys 04, Baltimore, USA, 2004, pp. 5061.
Knowl. Eng. Rev. 19 (2004) 213233. [70] S. Ray, R. Ungrangsi, F.D. Pellegrini, A. Trachtenberg, D.
[43] V. Andrikopoulos, S. Benbernou, M.P. Papazoglou, Managing the Starobinski, Robust location detection in emergency sensor
evolution of service specications, in: Proceedings of 20th networks, in: Proceedings of INFOCOM, S. Francisco, CA, 2003,
International Conference on the Advanced Information Systems pp. 10441053.
Engineering, CAiSE 2008, Montpellier, France, 2008, pp. 359374. [71] S. Gezici, Z. Tian, G.B. Giannakis, H. Kobayashi, A.F. Molisch, H.V.
[44] M.P. Papazoglou, W. van den Heuvel, Service oriented Poor, Z. Sahinoglu, Localization via ultra-wideband radios, IEEE
architectures: approaches, technologies and research issues, VLDB Signal Process. Mag. 22 (2005) 7084.
J. 16 (2007) 389415. [72] C. Decker et al., Costbenet model for smart items in the supply
[45] M. Papazoglou, Foresight & research priorities for service oriented chain, in: Proceedings of IOT Conference, Zurich, Switzerland, 2008,
computing, in: Proceedings of the 11th International Conference on pp. 155172.
Enterprise Information Systems, Milan, Italy, 2009, pp. 56. [73] C. Bornhvd, T. Lin, S. Haller, J. Schaper, Integrating automatic data
[46] T. Yu, M. Winslett, A unied scheme for resource protection in acquisition with business processes experiences with saps auto-id
automated trust negotiation, in: Proceedings of IEEE International infrastructure, in: VLDB 04: Proceedings of the Thirtieth
Symposium Security and Privacy, Colorado Springs, 2003, pp. 110 International Conference on Very Large Data Bases, VLDB
122. Endowment, 2004, pp. 11821188.
[47] Perci (pervasive service interaction). <http://www.hcilab.org/ [74] A.S. Tanenbaum, M.V. Steen, Distributed Systems: Principles and
projects/perci/index.html>. Paradigms, Prentice Hall PTR, Upper Saddle River, NJ, USA, 2001.
[48] W. Ye, J. Heidemann, D. Estrin, An energy-efcient MAC protocol for [75] J.N. Tsitsiklis, Problems in decentralized decision making and
wireless sensor networks, in: Proceedings of IEEE INFOCOM, vol. 3, computation, Ph.D. thesis, Massachusetts Institute of Technology,
2002, pp. 15671576. 1984.
[49] J. Polastre, J. Hill, D. Culler, Versatile low power media access for [76] M. Jelasity, A. Montresor, O. Babaoglu, Gossip-based aggregation in
wireless sensor networks, in: Proceedings of ACM SenSys, large dynamic networks, ACM Trans. Comput. Syst. 23 (1) (2005)
Baltimore, MD, USA, 2004, pp. 95107. 219252.
[50] G. Lu, B. Krishnamachari, C. Raghavendra, An adaptive energy- [77] D. Kempe, A. Dobra, J. Gehrke, Gossip-based computation of
efcient and low-latency MAC for data gathering in sensor aggregate information, in: Proceedings of FOCS, IEEE, Cambridge,
networks, in: Proceedings of International Workshop on MA, USA, 2003, pp. 482491.
Algorithms for Wireless, Mobile, Ad Hoc and Sensor Networks [78] J. Garay, Y. Moses, Fully polynomial byzantine agreement for n > 3t
(WMAN), Santa Fe, NM, 2004, p. 224. processors in t + 1 rounds, SIAM J. Comput. 27 (1) (1998) 247
[51] C. Enz, A. El-Hoiydi, J.-D. Decotignie, V. Peiris, WiseNET: an 290.
ultralow-power wireless sensor network solution, IEEE Comput. [79] D. Spanos, R. Olfati-Saber, R.M. Murray, Distributed sensor fusion
37 (8) (2004) 6270. using dynamic consensus, in: Proceedings of IFAC05, Prague, 2005,
[52] C. Schurgers, V. Tsiatsis, S. Ganeriwal, M. Srivastava, Optimizing p. 99.
sensor networks in the energy-latency-density design space, IEEE [80] L. Xiao, S. Boyd, S. Lall, A scheme for asynchronous distributed
Trans. Mob. Comput. 1 (1) (2002) 7080. sensor fusion based on average consensus, in: Proceedings of
[53] G. Merrett, N. White, N. Harris, B. Al-Hashimi, Energy-aware IPSN05, Los Angeles, USA, 2005, p. 99.
simulation for wireless sensor networks, in: Proceedings of IEEE [81] T.H. Cormen, C.E. Leiserson, R.L. Rivest, C. Stein, Introduction to
SECON, Rome, Italy, 2009, pp. 6471. Algorithms, MIT Press, 2001.
[54] A. Kansal, J. Hsu, S. Zahedi, M.B. Srivastava, Power management in [82] R.C. Shah, S. Roy, S. Jain, W. Brunette, Data mules: modeling a three-
energy harvesting sensor networks, ACM Trans. Embed. Comput. tier architecture for sparse sensor networks, in: Proceedings of IEEE
Syst. 6 (4) (2007) 32. International Workshop on Sensor Networks Protocols and
[55] A. Giridhar, P.R. Kumar, Computing and communicating functions Applications, 2003, pp. 3041.
over sensor networks, IEEE JSAC 23 (4) (2005) 755764. [83] F. De Pellegrini, C. Moiso, D. Miorandi, I. Chlamtac, R-P2P: a data
[56] P. Gupta, P.R. Kumar, The capacity of wireless networks, IEEE Trans. centric dtn middleware with interconnected throwboxes, in:
Inform. Theory 46 (2) (2000) 388404. Proceedings of Autonomics, Turin, Italy, 2008, pp. 1120.
[57] A.R.I. Schizas, G.B. Giannakis, Distributed estimation with ad hoc [84] I. Carreras, I. Chlamtac, F. De Pellegrini, D. Miorandi, BIONETS: bio-
wireless sensor networks, in: Proceedings of EURASIP EUSIPCO, inspired networking for pervasive communication environments,
Florence, 2006. IEEE Trans. Veh. Technol. 56 (1) (2007) 218229.
[85] J. Padhye, V. Firoiu, D. Towsley, J. Kurose, Modeling TCP throughput: [113] M. Bagaa, N. Lasla, A. Ouadjaout, Y. Challal, SEDAN: secure and
a simple model and its empirical validation, in: Proceedings of ACM efcient protocol for data aggregation in wireless sensor networks,
SIGCOMM, Vancouver, CA, 1998, pp. 303314. in: Proceedings of IEEE LCN, Dublin, Ireland, 2007, pp. 10531060.
[86] I. Koren, C.M. Krishna, Fault Tolerant Systems, Morgan Kaufman [114] C. Castelluccia, E. Mykletun, G. Tsudik, Efcient aggregation of
Publishers Inc., San Francisco, CA, USA, 2007. encrypted data in wireless sensor networks, in: Proceedings of
[87] R.S. Leslie Lamport, M. Pease, The byzantine generals problem, ACM MobiQuitous, San Diego, CA, USA, 2005, pp. 109117.
Trans. Program. Lang. Syst. 4 (3) (1982) 382401. [115] J. Girao, D. Westhoff, M. Schneider, CDA: concealed data
[88] S. Dolev, Self-stabilization, MIT Press, Cambridge, MA, USA, 2000. aggregation for reverse multicast trafc in wireless sensor
[89] W. Masri, Z. Mammeri, Middleware for wireless sensor networks: a networks, in: Proceedings of IEEE ICC, Seoul, Korea, 2005, pp.
comparative analysis, in: Proceedings of IFIP International Network 30443049.
and Parallel Computing Workshops (NPC), China, 2007, pp. 349356. [116] R. Riggio, S. Sicari, Secure aggregation in hybrid mesh/sensor
[90] J. Charles, Middleware moves to the forefront, Computer 32 (5) networks, in: Proceedings of SASN, Saint Petersburg, Russia, 2009,
(1999) 1719. pp. 16.
[91] N. Ibrahim, Orthogonal classication of middleware technologies, [117] A. Coen-Porisini, S. Sicari, SeDAP: Secure data aggregation protocol
in: Proceedings of UBICOMM 09, Sliema, Malta, 2009, pp. 4651. in privacy aware wireless sensor networks, in: Springer
[92] Object Naming Service (ONS) Standard. <http:// Proceedings of the 2nd International Conference on Sensor
www.epcglobalinc.org/standards/ons/ons101-standard- Systems and Software, Miami, Florida, USA, 2010.
20080529.pdf>. [118] E. Mykletun, J. Girao, D. Westhoff, Public key based cryptoschemes
[93] G. Weiss, Multiagent Systems. A Modern Approach to Distributed for data concealment in wireless sensor networks, in: Proceedings
Modern Approach to Articial Intelligence, MIT Press, Cambridge, of IEEE ICC, Istanbul, Turkey, 2006, pp. 22882295.
MA, USA, 1999. [119] L. Eschenauer, V.D. Gligor, A key-management scheme for
[94] T.W. Sandholm, V.R. Lesser, Coalitions among computationally distributed sensor networks, in: Proceedings of ACM CCS,
bounded agents, Artif. Intell. 1 (94) (1997) 99137. Washington, DC, USA, 2002, pp. 4147.
[95] D. Fudenberg, J. Tirole, Game Theory, MIT Press, 1991. [120] R.D. Pietro, A. Mei, L.V. Mancini, Random key assignment for secure
[96] D. Estrin, D. Culler, K. Pister, G. Sukhatme, Connecting the physical wireless sensor networks, in: Proceedings of ACM SASN, Fairfax,
world with pervasive networks, IEEE Pervasive Comput. 1 (1) VA, USA, 2003, pp. 6271.
(2002) 5969. [121] R.D. Pietro, C. Soriente, A. Spognardi, G. Tsudik, Collaborative
[97] W. Weber, J. Rabaaey, E. Aarts (Eds.), Ambient Intelligence, authentication in unattended WSNs, in: Proceedings of ACN WiSec,
Springer-Verlag, Berlin, 2005. Zurich, Switzerland, 2009, pp. 237244.
[98] E. Zelkha, B. Epstein, From devices to ambient intelligence, in: [122] T. Mielikinen, Privacy problems with anonymized transaction
Digital Living Room Conference, Laguna Niguel, California, 1998. databases, in: Proceedings of international Conference on
[99] M. Berger, F. Fuchs, M. Pirker, Ambient intelligence from personal Discovery Science (DS 2004), Lecture Notes in Computer Science,
assistance to intelligent megacities, in: Proceedings of Conference vol. 3245, Springer, 2004.
on Advances in Ambient Intelligence, 2007, pp. 2135. [123] A. Narayanan, V. Shmatikov, Obfuscated databases and group
[100] M. Kranz, P. Holleis, A. Schmidt, Embedded interaction: interacting privacy, in: Proceedings of ACM International Conference on
with the internet of things, IEEE Internet Comput. 14 (2010) 4653. Computer and Communications Security (CCS), ACM Press, New
[101] R. Sandhu, E. Coyne, H. Feinstein, C. Youman, Role-based access York, USA, 2005, pp. 102111.
control models, IEEE Comput. 29 (2) (1996) 3847. [124] P. Samarati, L. Sweeney, Protecting privacy when disclosing
[102] S. Papadopoulos, Y. Yang, D. Papadias, CADS: continuous information: k-anonymity and its enforcement through
authentication on data streams, in: Proceedings of the 33rd generalization and suppression, Technical Report SRI-CSL-98-04,
International Conference on Very Large Data Base (VLDB07), Computer Science Laboratory, SRI International.
Morgan Kaufmann Publishers Inc., Vienna, Austria, 2007, pp. [125] A. Bhargavspantzel, A. Squicciarini, E. Bertino, Trust negotiation in
135146. identity management, IEEE Secur. Priv. (2007) 5563.
[103] M. Ali, M. ElTabakh, C. Nita-Rotaru, Ft-Rc4: A Robust Security [126] A.V. Lamsweerde, E. Handling, Obstacles in goal-oriented
Mechanism for Data Stream Systems, Purdue University, Technical requirement engineering, IEEE Trans. Softw. Eng. 26 (2000) 978
Report, TR-05-024. 1005.
[104] W. Lindner, J. Meier, Securing the borealis data stream engine, in: [127] L. Liu, E. Yu, J. Mylopoulos, Analyzing security requirements as
Proceedings of the International Database Engineering and relationships among strategic actors, in: Proceedings of
Application Symposium (IDEAS06), Delhi, India, 2006, pp. 137 International Symposium on Requirements Engineering for
147. Information Security (SREIS), Raleigh, North Carolina, USA, 2002.
[105] R. Nehme, E. Rundesteiner, E. Bertino, A security punctuation [128] H. Mouratidis, P. Giorgini, G.A. Mason, Integrating security and
framework for enforcing access control on streaming data, in: systems engineering: towards the modelling of secure information
Proceedings of the 24th International Conference on Data system, in: Proceedings of International Conference on Advanced
Engineering (ICDE08), Cancun, Mexico, 2008, pp. 406415. Information System Engineering (CAiSE), Lecture Notes in
[106] M. Hammad, M. Franklin, W. Aref, A. Elmagarmid, Scheduling the Computer Science, vol. 2681, Springer, 2003, pp. 6378.
shared window joins over data streams, in: Proceedings of the 29th [129] L. Chung, Dealing with security requirements during the
International Conference on Very Large Data Base (VLDB03), development of information system, in: Proceedings of
Morgan Kaufman Publishers Inc., Berlin, Germany, 2003, pp. 297 International Conference on Advanced Information System
308. Engineering (CAiSE), vol. Klagenfurt/Velden, Austria, Lecture
[107] B. Carminati, E. Ferrari, K. Tan, J. Cao, A framework to enforce access Notes in Computer Science, Springer, Paris, France, 1993, pp.
control over data streams, ACM Trans. Inform. Syst. Sec. (TISSEC) 13 234251.
(3) (2008) 131. [130] J. Mylopolulos, L. Chung, B. Nixon, Representing and using non
[108] B. Carminati, E. Ferrari, K. Tan, Enforcing access control policies on functional requirements: a process oriented approach, IEEE Trans.
data streams, in: Proceedings of the 12th ACM Symposium on Softw. Eng. 18 (1998) 483497.
Access Control Models and Technologies (SACMAT07), Sophia [131] A. Anton, Goal-based requirements analysis, in: Proceedings of IEEE
Antipolis, France, 2007, pp. 2130. International Conference on Requirements Engineering (ICRE 96),
[109] B. Carminati, E. Ferrari, K. Tan, Specifying access control policies on Colorado Springs, 1996, pp. 136144.
data streams, in: Proceedings of the 12th International Conference [132] E. Kavakli, C. Kalloniatis, P. Loucopoulos, S. Gritzalis, Addressing
on Database System for Advanced Applications, Lecture Notes in privacy requirements in system design: the pris method, J.
Computer Science, Springer, Bangkok, Thailand, 2007, pp. 410421. Requirements Eng. 13 (3) (2008) 241255.
[110] L. Hu, D. Evans, Secure data aggregation in wireless sensor [133] A. Coen-Porisini, P. Colombo, S. Sicari, Privacy aware systems: from
networks, in: Proceedings of IEEE WSAAN, Orlando, Florida, USA, models to patterns, in: Software Engineering for Secure Systems:
2003, pp. 93105. Industrial and Research Perspectives, IGI Global, 2010.
[111] A. Mahimkar, T. Rappaport, SECUREDAV: a secure data aggregation [134] M. Blaze, J. Feigenbaum, J. Lacy, Decentralized trust management,
and verication protocol for sensor networks, in: Proceedings of in: Proceedings of IEEE International Symposium Security and
IEEE Globecom, Dallas, Texas, USA, 2004, pp. 21752179. Privacy, Colorado Springs, 1996, pp. 164173.
[112] B. Przydatek, D. Song, A. Perrig, SIA: secure information aggregation [135] K. Ren, T. Li, Z. Wan, F. Bao, R. Deng, K. Kim, Highly realible trust
in sensor networks, in: Proceedings of ACM SenSys, Los Angeles, establishment scheme in ad hoc networks, Comput. Netw. 45 (6)
California, USA, 2003, pp. 255265. (2004) 687699.
[136] Z. Liang, W. Shi, Enforcing cooperative resource sharing in Commitee member of S-Cube 2012. She is a TPC member of the inter-
untrusted peer to peer environment, ACM J. Mobile Netw. Appl. national conference IWCMC 2012, SENSORNETS 2012, S-Cube 2012, ICDT
10 (6) (2005) 771783. 2012, Mobility 2012, SNDS 2012, Q2SWinet 2012
[137] Y.-B. Lin, I. Chlamtac, Wireless and Mobile Network Architectures,
John Wiley & Sons, Inc., New York, NY, USA, 2000.
Francesco De Pellegrini received the Laurea
degree in 2000 and the Ph.D. degree in 2004,
Daniele Miorandi is the head of the iNSPIRE both in Telecommunication Engineering, from
Area at CREATE-NET, Italy. He received a PhD the University of Padova. During year 2001/
in Communications Engineering from Univ. of 2002 he spent one year at Boston University
Padova, Italy, in 2005, and a Laurea degree as a visiting scholar. He is currently a senior
(summa cum lauda) in Communications researcher and Deputy Area Head of the
Engineering from Univ. of Padova, Italy, in iNspire group at CREATE-NET. His research
2001. He joined CREATE-NET in Jan. 2005, interests are location detection, multirate
where he is leading the iNSPIRE (Networking systems, routing, wireless mesh networks,
and Security Solutions for Pervasive Comput- VoIP, Ad Hoc and Delay Tolerant Networks. F.
ing Systems: Research & Experimentation). De Pellegrini has been a TPC member of IEEE
His research interests include bio-inspired Infocom and a reviewer for several international networking conferences
approaches to networking and service provi- and journals. Francesco serves in the Steering Programm Commitee of
sioning in large-scale computing systems, modeling and performance Mobiquitous and Complex Conferences. Francesco was the Vice-chair for
evaluation of wireless networks, prototyping of wireless mesh solutions. the rst edition of Robocomm.
Dr. Miorandi has co-authored more than 100 papers in internationally
refereed journals and conferences. He serves on the Steering Committee
of various international events (WiOpt, Autonomics, ValueTools), for Imrich Chlamtac is the President of CREATE-
some of which he was a co-founder (Autonomics and ValueTools). He also NET and the Bruno Kessler Professor at the
serves on the TPC of leading conferences in the networking eld, University of Trento, Italy and has held vari-
including, e.g., IEEE INFOCOM, IEEE ICC, IEEE Globecom. He is a member of ous honorary and chaired professorships in
IEEE, ACM and ICST. USA and Europe including the Distinguished
Chair in Telecommunications Professorship at
the University of Texas at Dallas, Sackler
Sabrina Sicari was born on September 18, Professorship at Tel Aviv University and Uni-
1977 in Catania, Sicily, Italy. She received her versity Professorship at the Tech- nical Uni-
laurea degree in Electronical Engineering,110/ versity of Budapest. In the past he was with
110 cum laude, from University of Catania, Technion and UMass, Amherst, DEC Research.
Catania, Italy, in 2002. In March 2006 she got Dr. Imrich Chlamtac has made signicant
her Ph.D. in Computer and Telecommunica- contribution to various networking technologies as scientist, educator
tions Engineering at the same university. and entrepreneur. Dr. Chlamtac is the recipient of multiple awards and
From September 2004 to March 2006 she has recognitions including Fellow of the IEEE, Fellow of the ACM, Fulbright
been a Visiting Scholar at Dipartimento di Scholar, the ACM Award for Outstanding Contributions to Research on
Elettronica e Informatica, Politecnico di Mobility and the IEEE Award for Outstanding Technical Contributions to
Milano, Italy. Since May 2006 she works at Wireless Personal Communications. Dr. Chlamtac published close to four
Dipartimento di Informatica e Comunicazi- hundred refereed journal, book, and conference articles and is listed
one, Universit degli Studi dellInsubria in software engineering group among ISIs Highly Cited Researchers in Computer Science. Dr. Chlamtac is
(head Prof. Alberto Coen-Porisini). Dr. Sicari is an IEEE member. She is the co-author of four books, including the rst book on Local Area Net-
reviewer of Pervasive and Mobile Computing (Elsevier), IEEE Transactions works (1980) and the Amazon.com best seller and IEEE Editors Choice
on Vehicular Technology, ACM-Monet, International Journal of Computer Wireless and Mobile Network Architectures, published by John Wiley and
Applications in Technology (IJCAT), IEEE ICC09, IEEE ICC10, IEEE ISIE10, Sons (2000). Dr. Chlamtac has widely contributed to the scientic com-
S-Cube09, WiOpt09, Mobility11, ICDT 2012,and TPC member of munity as founder and Chair of ACM Sigmobile, founder and steering
Q2SWinet 2011, Q2SWinet 2010, Q2SWinet 2009, IEEE Globecom11, IEEE committee chair of some of the lead conferences in net-working,
Globecom10, IEEE ICC11, GIIS11, IWCMC 11, the international work- including ACM Mobicom, IEEE/SPIE/ACM OptiComm, CreateNet Mobiq-
shop SESENA 2010 (co-located with ICSE10), S-cube10. Dr. Sicari has uitous, CreateNet WiOpt, IEEE/CreateNet Broadnet, IEEE/CreateNet Tri-
been the general co-chair of S-Cube 2009. She has been a Steering Com- dentcom and IEEE/CreateNet Securecomm conferences. Dr. Chlamtac also
mittee member of S-Cube 2010 and guest editor for the ACM Monet serves as the founding Editor in Chief of the ACM/URSI/Springer Wireless
Special Issue, named Sensor, system and Software. Dr. Sicari is an Editor Networks (WINET), the ACM/Springer Journal on Special Topics in Mobile
for Computer Networks (Elsevier) journal since 2008. She is a Steering Networks and Applications (MONET).

Internet of Things - Vision, Applications and Research Challenges

Uploaded by

Copyright:

Available Formats

Internet of Things - Vision, Applications and Research Challenges

Uploaded by

Document Information

Original Description:

Original Title

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

Internet of Things - Vision, Applications and Research Challenges

Uploaded by

Copyright:

Available Formats

Ad Hoc Networks 10 (2012) 14971516

Contents lists available at SciVerse ScienceDirect

Internet of things: Vision, applications and research challenges

1. Introduction In such a perspective, the conventional concept of the

Fig. 1. Taxonomy of research areas relevant to Internet-of-Things.

5. Security denition of an access control mechanism and second,

Data stream access control,

General privacy model for

Fig. 2. Graphical representation of security challenges in Internet-of-Things.

model privacy requirements and a set of rules to transform 5.3. Trust

You might also like