A revised version of this manuscript is currently under review for possible publication in IEEE Wireless Communications

Long-Range Communications in Unlicensed Bands:

the Rising Stars in the IoT and Smart City Scenarios

arXiv:1510.00620v1 [cs.NI] 2 Oct 2015

Marco Centenaro, Student Member, IEEE, Lorenzo Vangelista, Senior Member, IEEE,
Andrea Zanella, Senior Member, IEEE, and Michele Zorzi, Fellow, IEEE

AbstractConnectivity is probably the most basic building

block of the Internet of Things (IoT) paradigm. Up to know,
the two main approaches to provide data access to the things
have been based either on multi-hop mesh networks using shortrange communication technologies in the unlicensed spectrum,
or on long-range, legacy cellular technologies, mainly 2G/GSM,
operating in the corresponding licensed frequency bands. Recently, these reference models have been challenged by a new type
of wireless connectivity, characterized by low-rate, long-range
transmission technologies in the unlicensed sub-GHz frequency
bands, used to realize access networks with star topology which
are referred to a Low-Power Wide Area Networks (LPWANs).
In this paper, we introduce this new approach to provide
connectivity in the IoT scenario, discussing its advantages over
the established paradigms in terms of efficiency, effectiveness, and
architectural design, in particular for the typical Smart Cities
applications.
Index TermsInternet of Things, Smart Cities, Low-Power
TM
TM
TM
Wide Area Network (LPWAN), LoRa , SIGFOX , Ingenu ,
Cellular IoT.

I. I NTRODUCTION
The Internet of Things (IoT) paradigm refers to a network
of interconnected things. The network is normally intended
as the IP network and the things are devices, such as sensors
and/or actuators, equipped with a telecommunication interface
and with processing and storage units. This communication
paradigm should hence enable seamless integration of potentially any object into the Internet, thus allowing for new forms
of interactions between human beings and devices, or directly
between device and device, according to what is commonly
referred to as the Machine-to-Machine (M2M) communication
paradigm [1].
The development of the IoT is an extremely challenging
topic and the debate on how to put it into practise is still
open. The discussion interests all layers of the protocol stack,
from the physical transmission up to data representation and
service composition. However, the whole IoT castle rests on
the wireless technologies that are used to provide data access
to the end devices.
The authors are with the Department of Information Engineering, University of Padova, Italy; e-mail: {firstname.lastname}@dei.unipd.it.
L. Vangelista and M. Zorzi are also with Patavina Technologies s.r.l.,
Padova, Italy; e-mail: {firstname.lastname}@patavinatech.com,
web: http://www.patavinatech.com/en/.
This paper is partly based on the paper Long-range IoT technologies: the
TM
dawn of LoRa by L. Vangelista, A. Zanella and M. Zorzi presented at
Fabulous 2015 conference, Sept. 2325, 2015, Ohrid, Republic of Macedonia.

For many years, multi-hop short-range transmission technologies, such as ZigBee and Bluetooth, have been considered
a viable way to implement IoT services [2][4]. Although these
standards provide very low power consumption, which is a
fundamental requirement for many IoT devices like, e.g., smart
sensors, their limited coverage constitutes a major obstacle,
in particular when the application scenario involves services
that require urban-wide coverage, as in typical Smart City
applications [4]. The experimentations of some initial Smart
Cities services have, indeed, revealed the limits of the multihop short-range paradigm for this type of IoT applications,
stressing the need for an access technology that can allow for
a place-&-play type of connectivity, i.e., that makes it possible
to connect any device to the IoT by simply placing it in the
desired location and switching it on [5].
In this perspective, wireless cellular networks may play a
fundamental role in the spread of IoT, since they are able to
provide ubiquitous and transparent coverage [1], [6], [7]. In
particular, the Third Generation Partnership Project (3GPP),
which is the standardization body for the most important
cellular technologies, is attempting to revamp 2G/GSM to
support IoT traffic, implementing the so-called Cellular IoT
(CIoT) architecture [8]. On the other side, the latest cellular
network standards, e.g., UMTS and LTE, were not designed to
provide machine-type services to a massive number of devices.
In fact, differently from traditional broadband services, IoT
communication is expected to generate, in most cases, sporadic
transmissions of short packets. At the same time, the potentially
huge number of IoT devices asking for connectivity through a
single Base Station (BS) would raise new issues related to the
signaling and control traffic, which may become the bottleneck
of the system [5]. All these aspects make current cellular
network technologies not suitable to support the envisioned
IoT scenarios, while, on the other hand, a number of research
challenges still need to be addressed before the upcoming 5G
cellular networks may natively support IoT services.
A promising alternative solution, standing in between shortrange multi-hop technologies operating in the unlicensed
industrial, scientific, and medical (ISM) frequency bands, and
long-range cellular-based solutions using licensed broadband
cellular standards, is provided by the so-called Low-Power
Wide Area Networks (LPWANs).
These kinds of networks exploit sub-GHz, unlicensed
frequency bands and are characterized by long-range radio
links and star topologies. The end devices, indeed, are directly
connected to a unique collector node, generally referred to as
gateway, which also provides the bridging to the IP world. The

architecture of these networks is designed to provide wide area

coverage and ensure the connectivity also to nodes that are
deployed in very harsh environments.
The goal of this paper is to provide an introductory overview
of the LPWAN paradigm and of its main technological
interpretations. We will discuss the advantages provided by this
new type of connectivity with respect to the more traditional
solutions operating in the unlicensed spectrum, especially
for applications related to Smart Cities. To substantiate our
argumentation, we will refer to some preliminary experiments
TM
and deployments of IoT networks based on LoRa , one of
the LPWAN solutions available on the market today.
The rest of the paper is organized as follows. In Section II,
current wireless technologies and service platforms for the
IoT connectivity are reviewed. The potential of LPWANs
is discussed is Section III, while Section IV describes the
commercial LPWAN products available today, focusing in
TM
greater detail on LoRa , whose characteristics make it a
good representative of the LPWAN family, while its open
specifications make it possible to access some details of its most
interesting and specific mechanisms. In Section V we discuss
the experience gained with some experimental deployments
TM
of a LoRa network. Conclusions and final remarks can be
found in Section VI.
II. A Q UICK OVERVIEW OF C URRENT I OT
C OMMUNICATION S TANDARDS
Although the IoT paradigm does not set any constraint on
the type of technology used to connect the end devices to the
Internet, it is a fact that wireless communication is the only
feasible solution for a large majority of the IoT applications
and services. As mentioned, the current practice considers
either cellular-based or multi-hop short-range technologies. In
the latter case, the connected things usually run on dedicated
protocol stacks, suitably designed to cope with the constraints
of the end devices. Furthermore, at least one such device is
required to be connected to the IP network, acting as gateway
for the other nodes. The architecture is hence distributed, with
many islands (sub-nets) that operate according to different
connectivity protocols, and are connected to the IP network via
gateways. The applications and services are deployed on top of
this connectivity level, according to a distributed service layer.
The applications may run either locally, i.e., in the sub-net, or,
more and more often (as typical in the Smart City scenario),
using cloud computing services.
At this level we can find the IoT platforms that act as a
unifying framework, enabling the service creation and delivery,
as well as the operation, administration, and maintenance of
the things and the gateways. Nowadays, the most important
de facto standards in the IoT arena are the following:
1) extremely short-range systems, e.g., Near Field Communications (NFC) enabled devices;
2) short-range passive and active Radio Frequency IDentification (RFID) systems;
3) systems based on the family of IEEE 802.15.4 standards
TM
like ZigBee , 6LoWPAN, Thread-based systems;
4) Bluetooth-based systems, including Bluetooth Low Energy (BLE);

TM

TM

5) proprietary systems, including Z-Wave , CSRMesh ,

i.e., the Bluetooth mesh by Cambridge Silicon Radio (a
TM
company now owned by Qualcomm), EnOcean ;
TM
6) systems mainly based on IEEE 802.11/Wi-Fi , e.g.,
those defined by the AllSeen Alliance specifications,
which explicitly include the gateways, or by the Open
Interconnect Consortium.
The vast majority of the connected things at the moment
TM
is using IEEE 802.15.4-based systems, in particular ZigBee .
The most prominent features of these networks are that
they operate mainly in the 2.4 GHz and optionally in the
868/915 MHz unlicensed frequency bands and the network
level connecting these nodes1 uses a mesh topology. The
distances between the nodes in this kind of systems ranges
from few meters, up to roughly 100 meters, depending on the
surrounding environment (presence of walls, obstacles, and so
on).
To better appreciate the comparison with LPWAN technologies, it is worth highlighting the main characteristics of these
IoT technologies.
Mesh networking. Multihop communication is necessary to
extend the network coverage beyond the limited reach of
the low-power transmission technology used. Furthermore,
the mesh architecture can provide resilience to the failure
of some nodes. On the other hand, the maintenance of the
mesh network requires non-negligible control traffic, and
multi-hop routing generally yields long communication
delays, and unequal and unpredictable energy consumption
among the devices.
Short coverage range - high data rate. The link level
technologies used in these systems tend to privilege the
data rate rather than the sensitivity, i.e., in order to recover
from the network delays due to the mesh networking, these
networks have a relatively high raw link bit rate (e.g.,
250 Kbit/s), but they are not robust enough to penetrate
building walls and other obstacles (even in the 868/915
MHz band). In other words, in the trade-off between rate
and sensitivity, the rate is usually preferred.
III. A N EW PARADIGM : L ONG -R ANGE I OT
C OMMUNICATIONS IN U NLICENSED BANDS
As a counterpart of the unlicensed short-range technologies
for the IoT mentioned in the previous sections, we turn our
attention to the emerging paradigm of LPWAN.
Most LPWANs operate in the unlicensed ISM bands centered
at 2.4 GHz, 868/915 MHz, 433 MHz, and 169 MHz, depending
on the region of operation. The radio emitters operating in
these frequency bands are commonly referred to as Short
Range Devices [9], a rather generic term that delivers the
idea of coverage ranges of few meters, which was indeed
the case for the previous ISM wireless systems. Nonetheless,
the ERC Recommendation 70-03 specifies that The term
Short Range Device (SRD) is intended to cover the radio
transmitters which provide either uni-directional or bidirectional communication which have low capability of
1 Node is a term that is frequently used to indicate a connected thing, with
emphasis on the communication part.

causing interference to other radio equipment. Therefore, cellular systems architecture that, however, is stripped of
there is no explicit mention of the actual coverage range of most advanced features, such as the management of user
such technologies.
mobility and resource scheduling. The combination of the
LPWAN solutions are indeed examples of short-range simple but effective topology of cellular systems with a
devices with cellular-like coverage ranges, in the order of much lighter management plane, makes the LPWAN approach
1015 km in rural areas, and 25 km in urban areas. This is particularly suitable to support services with relatively low
possible thanks to a radically new physical layer design, aimed Average Revenue Per User, such as those envisioned in the
at very high receiver sensitivity. For example, while the nominal Smart City scenario.
TM
sensitivity of ZigBee and Bluetooth receivers is about -125
dBm and -90 dBm, respectively, the typical sensitivity of a
IV. A R EVIEW OF L ONG -R ANGE I OT C OMMUNICATIONS
LPWAN receiver is around -150 dBm (see Section IV).
S YSTEMS IN U NLICENSED BANDS
The downside of these long-range connections is the low
In this section we quickly overview three of the most
TM
data rate, which usually ranges from few hundred to few
prominent technologies for LPWANs, namely SIGFOX ,
TM
TM
thousand bit/s, significantly lower than the bitrates supported
Ingenu , and LoRa . In particular, we will describe in
TM
by the actual short-range technologies, e.g., 250 Kbit/s in
TM
greater detail the LoRa technology, which is gaining more
ZigBee and 12 Mbit/s in Bluetooth. However, because of
and more momentum, and whose specifications are publicly
the signaling overhead and the multi-hop packet forwarding
available, thus making it possible to appreciate some of the
method, the actual flow-level throughput provided by such
technical choices that characterize LPWAN solutions. In Tab. I
short-range technologies is generally much lower than the
a comparison between these LPWAN radio technologies can
nominal link-layer bitrate, settling to values that are comparable
be found.
to those reached by the single-hop LPWANs. While such
low bitrates are clearly unsatisfactory for most common dataTM
hungry network applications, many Smart City and IoT services A. SIGFOX
TM
are expected to generate a completely different pattern of
SIGFOX ,2 the first LPWAN technology proposed in the
traffic, characterized by sporadic and intermittent transmissions IoT market, was founded in 2009 and has been growing very
TM
of very small packets (typical of monitoring and metering fast since then. The SIGFOX physical layer employs aa Ultra
applications, remote switching control of equipment, and so Narrow Band (UNB) wireless modulation, while the network
TM
on). Furthermore, many of these applications are rather tolerant layer protocols are the secret sauce of the SIGFOX network
to delays and packet losses and, hence, are suitable for the and, as such, there exists basically no publicly available
TM
connectivity service provided by LPWANs.
documentation. Indeed, the SIGFOX business model is that
Another important characteristic of LPWANs is that the of an operator for IoT services, which hence does not need to
things, i.e., the end devices, are connected directly to one open the specifications of its inner modules.
(or more) gateway with a single-hop link, very similar to the
The first releases of the technology only supported uniclassic cellular network topology. This greatly simplifies the directional uplink communication, i.e., from the device towards
coverage of large areas, even nation-wide, by re-using the the aggregator; however bi-directional communication is now
existing infrastructure of the cellular networks. For example, supported. SIGFOXTM claims that each gateway can handle
TM
LoRa systems are being deployed by telecommunication up to a million connected objects, with a coverage area of
operators like Orange and Bouygues Telecom in France, by 3050 km in rural areas and 310 km in urban areas.
Swisscom in Switzerland, and by KPN in the Netherlands,
TM
TM
while SIGFOX has already deployed a nation-wide access
B. Ingenu
network for M2M and IoT devices in many central European
TM
An emerging star in the landscape of LPWANs is Ingenu ,
countries, from Portugal to France. Furthermore, the star
topology of LPWANs makes it possible to have greater control a trademark of On-Ramp Wireless, a company headquartered in
3
of the connection latency, thus potentially enabling the support San Diego (USA). On-Ramp Wireless has been pioneering the
of interactive applications that require predictable response 802.15.4k standard [10]. The company developed and owns the
times such as, for example, the remote control of street lights rights of the patented technology called Random Phase Multiple
R

in a large city, the operation of barriers to limited-access streets, Access (RPMA ) [11], which is deployed in different networks.
Conversely to the other LPWAN solutions, this technology
the intelligent control of traffic lights, and so on.
Besides the access network, the similarity between LPWANs works in the 2.4 GHz band but, thanks to a robust physical
and legacy cellular systems further extends to the bridging of layer design, can still operate over long-range wireless links
the technology-specific wireless access to the IP-based packet and under the most challenging RF environments.
switching core network. Indeed, the LPWAN gateways play a
similar role as the Gateway GPRS Support Node (GGSN) in C. The LoRaTM System
TM
GPRS/UMTS networks, or the Evolved Packet Core in LTE,
LoRa is a new physical layer LPWAN solution, which
acting as point-of-access for the end devices to the IP-based
has been designed and patented by Semetch Corporation
core network and forwarding the data generated by things to a
2 http://www.sigfox.com
logic controller, usually named Network Server.
3 http://www.onrampwireless.com
Therefore, LPWANs inherit the basic aspects of the legacy

that also manufactures the chipsets [12]. More specifically,

the PHY is a derivative of Chirp Spread Spectrum (CSS)
[13], where the innovation consists in ensuring the phase
continuity between different chirp symbols in the preamble
part of the physical layer packet, thus enabling a simpler and
more accurate timing and frequency synchronization, without
requiring expensive components that generate a stable local
TM
TM
clock in the LoRa node.
Fig. 1: LoRa system architecture
The technology employs a spreading technique, according
to which a symbol is encoded in a longer sequence of bits,
thus reducing the signal to noise and interference ratio required
at the receiver for correct reception, without changing the where a channel is identified by a specific sub-band and
frequency bandwidth of the wireless signal. The length of the spreading factor.
This access mode greatly simplifies the management of the
spreading code can be varied, thus making it possible to provide
network
access for the end nodes, moving all the complexity
variable data rates, giving the possibility to trade throughput
to
the
NetServer.
Furthermore, the end nodes can freely move
for coverage range, or link robustness, or energy consumption.
across
cells
served
by different gateways without generating
The system has been designed to work in the 169 MHz,
any
additional
signaling
traffic in the access network, nor in the
433 MHz and 915 MHz bands in the USA, but in Europe
core
network.
Finally,
we
observe that increasing the number
it works in the 868 MHz band. According to the regulation
of
gateways
that
serve
a
certain
end device will increase the
in [14], the radio emitters are required to adopt duty cycled
reliability
of
its
connection
to
the
NetServer, which may be
transmission (1% or 0.1%, depending on the sub-band), or the
interesting
for
critical
applications.
so-called Listen Before Talk (LBT) Adaptive Frequency Agility
TM
A distinguishing feature of the LoRa network is that it
(AFA) technique, a sort of carrier sense mechanism used to
prevent severe interference among devices operating in the envisages three classes of end devices, named Class A (for All),
TM
same band. According to the specification in [15], LoRa (as Class B (for Beacon) and Class C (for Continuously listening),
TM
well as SIGFOX ) uses the duty cycled transmission option each associated to a different operating mode [15].
Class A defines the default functional mode of the
only, which limits the rate at which the end device can actually
TM
LoRa
networks, and must be mandatorily supported by all
generate messages. However, by supporting multiple channels,
TM
TM
LoRa
devices.
In a Class A network, transmissions are always
LoRa makes it possible for an end node to engage in longer
initiated
by
the
end
devices, in a totally asynchronous manner.
data exchange procedures by changing carrier frequency, while
After
each
uplink
transmission,
the end device will open (at
respecting the duty cycle limit in each channel.
TM
least)
two
reception
windows,
waiting
for any command or
While the PHY layer of LoRa is proprietary, the rest of
TM
data
packet
returned
by
the
NetServer.
The second window
the protocol stack, known as LoRaWAN , is kept open, and
TM
4
is
opened
on
a
different
sub-band
(previously
agreed with
its development is carried out by the LoRa Alliance, led by
the
NetServer)
in
order
to
increase
the
resilience
against
IBM, Actility, Semtech, and Microchip.
TM
channel
fluctuations.
Class
A
networks
are
mainly
intended
for
As exemplified in Fig. 1, the LoRa network is typically
monitoring
applications,
where
the
data
which
are
produced
laid out in a star-of-stars topology, where the end devices
TM
are connected via a single-hop LoRa link to one or many by the end devices have to be collected by a control station.
Class B has been introduced to decouple uplink and downlink
gateways that, in turn, are connected to a common Network
transmissions.
Class B end devices, indeed, synchronize with
Server (NetServer) via standard IP protocols.
the
NetServer
by
means of beacon packets which are broadcast
The gateways relay messages between the end devices and
by
Class
B
gateways
and can hence receive downlink data
the NetServer according to the protocol architecture represented
or
command
packets
in
specific time windows, irrespective
in Fig. 2. Conversely to standard cellular network systems,
of
the
uplink
traffic.
Therefore,
Class B is then intended for
however, the end devices are not required to associate to a
end
devices
that
need
to
receive
commands from a remote
certain gateway to get access to the network, but only to the
controller,
e.g.,
switches
or
actuators.
NetServer. The gateways act as a sort of relay/bridge and simply
Finally, Class C is defined for end devices without (strict)
forward to their associated NetServer all successfully decoded
energy
constraints (e.g., connected to the power grid), which
messages sent by any end device, after adding some information
can
hence
keep the receive window always open.
regarding the quality of the reception. The NetServer is hence
It
is
worth
noting that, at the time of writing, Class A and B
in charge of filtering duplicate and unwanted packets, and
specifications
are provided in [15], while Class C specifications
of replying to the end devices by choosing one of the inare
still
in
draft
form.
range gateways, according to some criterion (e.g., better radio
TM
The
MAC
layer,
according to LoRaWAN specification
connectivity). The gateways are thus totally transparent to the
end devices, which are logically connected directly to the [15], is basically an ALOHA protocol controlled primarily by
TM
NetServer. Note that current full-fledged LoRa Gateways the LoRa NetServer. A description of the protocol is beyond
TM
allow for the parallel processing of up to 9 LoRa channels, the scopeTMof this paper and can be found in [15]. Overall
the LoRa MAC has been designed attempting to mimic as
4 https://www.lora-alliance.org/
much as possible the IEEE 802.15.4 MAC. The objective is to
End-Device

End-Device

LoRaTM

End-Device

End-Device

End-Device

radio links

End-Device

LoRaTM

LoRaTM

Gateway

Gateway

IP connection

LoRaTM

End-Device

IP connection

NetServer

End-Device

End-Device

LoRaTM EndNode

LoRaTM NetServer

Application

Application

End-node
Side

Server
Side

LoRaTM
MAC
End-node side

LoRaTM
MAC
Server side

LoRaTM Gateway
Packet forwarder
LoRaTM
Radio link

LoRaTM
Physical Layer
chip SX127x

LoRaTM
Physical Layer
chip SX123x

Fig. 2: LoRa

TM

Backhaul
Link
(cellular, wired)

Legacy IP link

Backhaul
Link
(cellular, wired)

protocol architecture.

TM

simplify the accommodation, on top of the LoRa MAC, of

the major protocols now running on top of the IEEE 802.15.4
MAC, such as 6LoWPAN and CoAP. A clear analogy is the
authentication mechanism, which is taken directly from the
IEEE 802.15.4 standard using the 4-octet M IC (Message
Integrity Code).
V. S OME E XPERIMENTAL R ESULTS U SING A
TM
L O R A N ETWORK
In this section, we corroborate the argumentations of the
previous sections by reporting some observations based on
TM
some initial deployments of LoRa networks.
TM

A. A LoRa Deployment Test

TM

TM

Fig. 3: LoRa

gateway installation.

practical the LPWAN paradigm is for the Smart City scenario

and, on the other side, some intuition from the economical point
of view. Indeed, though extremely limited in its extent, the
positive experience gained in the proof-of-concept installation
TM
of the LoRa system in a building bodes well to the extension
of the service to other public and private buildings, realizing at
the same time an infrastructure for other Smart City services.
According to Analysis Mason 2014 data, indeed, the number
of LPWAN smart buildings connections is projected to be
0.8 billions by 2023; and according to the McKinsey Global
Institute analysis, the potential economic impact of IoT 2025
for Home and Cities is between $1,1 and $2,0 trillion. Thus,
LPWAN solutions appear to have both the technical and the
commercial capability to become the game changer in the
Smart City scenario.

A LoRa private network has been installed by Patavina

Technologies s.r.l. in a large and tall building (19 floors)
in Northern Italy, for a proof of concept of the capabilities
TM
of the LoRa network. The objective is to monitor and
control the temperature and the humidity of the different
TM
rooms, with the aim of reducing the costs related to heating, B. LoRa Coverage Analysis
ventilation, and air conditioning. To this end, different wireless
One of the most debated aspects of LPWAN is the actual
and wired communication technologies (including powerline coverage range. This is crucial for a correct estimation of
communication) have been tried, but these solutions have been the costs for city-wide coverage, which may clearly have an
mostly unsatisfactory, requiring the installation of repeaters and important impact on the Capital Expenditure of the service
gateways in basically every floor to guarantee mesh connectivity providers.
TM
and access to the IP backbone. Instead, the LoRa technology
To gain insight in this respect, we carried out a coverage
TM
has made it possible to provide the service by installing a experimental test of LoRa networks in the city of Padova,
single gateway on the ninth floor and placing 32 nodes all over Italy. The aim was to assess the worst case coverage of the
the building, at least one per floor. The installation included technology, to have a conservative estimate of the number of
the integration of the NetServer with a monitoring application gateways required to cover the whole city. To this end, we
and with the databases already in use. At the time of writing, placed a gateway with no antenna gain at the the top of a two
the installation has been flawlessly running for six months and storey building, without antenna elevation, in an area where
is being considered as the preferred technology for the actual high buildings are present.
implementation of the energy saving program in many other
Fig. 3 shows the experimental setup, while Fig. 4 shows the
buildings.
results of the test. It can be seen that, in such harsh propagation
TM
TM
We want to remark that the LoRa network connectivity conditions, the LoRa technology allows to cover a cell of
has been put under strain placing the nodes in elevators and in about 2 km of radius. However, the connection at the cell edge is
other places known to be challenging for radio connectivity. All guaranteed only when using the lowest bit rate (i.e., the longest
the stress tests have been successfully passed. The envisioned spreading sequence which provides maximum robustness), with
next step is to install a gateway on an elevated site to serve low margin for possible interference or to link budget changes.
multiple buildings in the neighborhood.
For this reason, we assumed a nominal coverage range of
This proof of concept is particularly relevant as it provides, 1.2 km, a value that ensures a reasonable margin to interference
on the one side, interesting insights on how pertinent and and link budget variations due, e.g., to fading phenomena.

TABLE I: Comparison between LPWAN radio technologies.

TM

Coverage range (km)

Frequency bands (MHz)
ISM band
Bi-directional link
Data rate (Kbps)
Nodes per BS

SIGFOX
rural: 3050
urban: 310
868 or 902
3
3
0.1
106

TM

TM

Ingenu

LoRa
rural: 1015
urban: 35
various, sub-GHz
3
3
0.337.5
104

15
2400
3
7
0.018
104

TM

Fig. 4: LoRa system single cell coverage in Padova, Italy.

Worst case test.
TM

Using this parameter, we attempted a rough coverage

planning for the city of Padova, which extends over an area
of about 100 square kilometers. The resulting plan is shown
in Fig. 5, from which we observe that, with the considered
conservative coverage range estimate, the coverage of the entire
municipality can be reached with a total of 30 gateways, which
is less than half the number of sites deployed by one of the
major cellular operators in Italy to provide mobile cellular
access over the same area.
Finally, we observe that Padova municipality accounts for
about 200000 inhabitants. Considering 30 gateways to cover
the city, we get about 7000 inhabitants per gateway. The current
TM
LoRa gateway technology claims the capability of serving
15000 nodes per gateway, which accounts for about 2 things
per person. Considering that the next generation of gateways
is expected to triple the capacity (by using multiple directional
antennas), in the long term we can expect that a basic coverage
of the city may grant up to 67 things per person, on average,
which seems to be more than adequate for most Smart City
applications. Any further increase in the traffic demand can be
addressed by installing additional gateways, a solution similar
to densification in cellular networks.
VI. C ONCLUSIONS
In this paper we have described the new emerging LowPower Wide Area Networks (LPWANs) paradigm for Internet
of Things connectivity. This solution is based on long-range
radio links, in the order of the tens of kilometers, and a star network topology, i.e., every node is directly connected to the base
station. Therefore, LPWANs are inherently different from usual
IoT architectures, which are, instead, typically characterized

Fig. 5: LoRa
case test.

system cell coverage for Padova, Italy. Worst

by short-range links and a mesh topology. The most prominent

TM
TM
TM
LPWAN technologies, i.e., SIGFOX , Ingenu , and LoRa ,
have been introduced and compared to the current short-range
communication standards. The experimental trials, which have
TM
been performed employing LoRa technology, have shown
that the LPWAN paradigm should be intended to complement
current IoT standards as an enabler of Smart City applications,
which can greatly benefit from long-range links.
ACKNOWLEDGMENT
The authors would like to thank Ivano Calabrese and Nicola
Bressan from Patavina Technologies, for their contributions in
TM
the implementation of the LoRa system and especially for
running the field trials.
R EFERENCES
[1] C. Anton-Haro and M. Dohler, Machine-to-Machine (M2M) Communications: Architecture, Performance and Applications, 1st ed. Woodhead
Publishing Ltd., Jan. 2015.
[2] J. Gubbi, R. Buyya, S. Marusic, and M. Palaniswami, Internet of Things
(IoT): A Vision, Architectural Elements, and Future Directions, Future
Generation Computer Systems, vol. 29, no. 7, pp. 16451660, Sept.
2013. [Online]. Available: http://dx.doi.org/10.1016/j.future.2013.01.010
[3] D. Miorandi, S. Sicari, F. De Pellegrini, and I. Chlamtac, Internet of
Things, Ad Hoc Networks, vol. 10, no. 7, pp. 14971516, Sept. 2012.
[Online]. Available: http://dx.doi.org/10.1016/j.adhoc.2012.02.016
[4] A. Zanella, N. Bui, A. Castellani, L. Vangelista, and M. Zorzi, Internet
of Things for Smart Cities, IEEE Internet of Things Journal, vol. 1,
no. 1, pp. 2232, Febr. 2014.

[5] A. Biral, M. Centenaro, A. Zanella, L. Vangelista, and M. Zorzi,

The challenges of M2M massive access in wireless cellular networks,
Digital Communications and Networks, vol. 1, no. 1, pp. 1 19, Feb.
2015. [Online]. Available: http://www.sciencedirect.com/science/article/
pii/S235286481500005X
[6] S.-Y. Lien, K.-C. Chen, and Y. Lin, Toward Ubiquitous Massive Accesses
in 3GPP Machine-to-Machine ommunications, IEEE Communications
Magazine, vol. 49, no. 4, pp. 6674, Apr. 2011.
[7] 3GPP, Standardization of Machine-Type Communications, Tech. Rep.
V0.2.2, Dec. 2013.
[8] Vodafone Group Plc., New Study Item on Cellular System Support for
Ultra Low Complexity and Low Throughput Internet of Things, 3GPP
TSG GERAN#62, Tech. Rep. GP-140421, May 2014.
[9] ETSI, Electromagnetic compatibility and Radio spectrum Matters
(ERM), Tech. Rep. 300 220-1 V2. 4.1, Jan. 2012.
[10] IEEE Standard for Local and metropolitan area networks Part 15.4:
Low-Rate Wireless Personal Area Networks (LR-WPANs) Amendment
5: Physical Layer Specifications for Low Energy, Critical Infrastructure
Monitoring Networks, IEEE Standard 802.15.4k-2013, pp. 1149, Aug.
2013.
[11] On-Ramp Wireless Inc., Light monitoring system using a random phase
multiple access system, July 2013, US Patent 8,477,830. [Online].
Available: https://www.google.com/patents/US8406275
[12] F. Sforza, Communications system, Mar. 2013, US Patent 8,406,275.
[Online]. Available: https://www.google.com/patents/US8406275
[13] A. J. Berni and W. Gregg, On the Utility of Chirp Modulation for Digital
Signaling, IEEE Transactions on Communications, vol. 21, no. 6, pp.
748751, June 1973.
[14] ETSI, Electromagnetic compatibility and Radio spectrum Matters
(ERM); Short Range Devices (SRD); Radio equipment to be used in the
25 MHz to 1 000 MHz frequency range with power levels ranging up
to 500 mW; Part 1: Technical characteristics and test methods, Tech.
Rep. EN 300 220-1 V2.4.1, Jan. 2012.
TM
TM
[15] LoRa Alliance, LoRa Specification V1.0, Tech. Rep., May 2015.

Marco Centenaro (S14) received the Bachelors degree in Information

Engineering in 2012 and the Masters degree in Telecommunication Engineering
in 2014, both from the University of Padova, Italy. Since November 2014 he is
a Ph.D student at the Department of Information Engineering of the University
of Padova, Italy. His research interests include the next generation of cellular
networks (5G) and in particular the Machine-to-Machine Communication.

Lorenzo Vangelista (S93-M97-SM02) received the Laurea and Ph.D.

degrees in electrical and telecommunication engineering from the University
of Padova, Padova, Italy, in 1992 and 1995, respectively. He subsequently
joined the Transmission and Optical Technology Department, CSELT, Torino,
Italy. From December 1996 to January 2002, he was with Telit Mobile
Terminals, Trieste, Italy, and then, until May 2003, he was with Microcell A/S,
Copenaghen, Denmark. In July 2006, he joined the Worldwide Organization of
Infineon Technologies as Program Manager. Since October 2006, he has been an
Associate Professor of Telecommunication with the Department of Information
Engineering, Padova University. His research interests include signal theory,
multicarrier modulation techniques, cellular networks and wireless sensors and
actuators networks.

Andrea Zanella (S98-M01-SM13) received the Laurea degree in computer

engineering and Ph.D. degree in electronic and telecommunications engineering
from the University of Padova, Padova, Italy, in 1998 and 2000, respectively. He
was a Visiting Scholar with the Department of Computer Science, University
of California, Los Angeles (UCLA), Los Angeles, CA, USA, in 2000. He
is an Assistant Professor with the Department of Information Engineering
(DEI), University of Padova. He is one of the coordinators of the SIGnals and
NETworking (SIGNET) research lab. His long-established research activities
are in the fields of protocol design, optimization, and performance evaluation
of wired and wireless networks.

Michele Zorzi (S89-M95-SM98-F07) received his Laurea and PhD degrees

in electrical engineering from the University of Padova in 1990 and 1994,
respectively. During academic year 1992-1993 he was on leave at UCSD,
working on multiple access in mobile radio networks. In 1993 he joined
the faculty of the Dipartimento di Elettronica e Informazione, Politecnico
di Milano, Italy. After spending three years with the Center for Wireless
Communications at UCSD, in 1998 he joined the School of Engineering
of the University of Ferrara, Italy, where he became a professor in 2000.
Since November 2003 he has been on the faculty of the Information
Engineering Department at the University of Padova. His present research
interests include performance evaluation in mobile communications systems,
random access in wireless networks, ad hoc and sensor networks, Internet-ofThings, energy constrained communications protocols, cognitive networks, and
underwater communications and networking. He was Editor-In-Chief of IEEE
Wireless Communications from 2003 to 2005 and Editor-In-Chief of the IEEE
Transactions on Communications from 2008 to 2011, and is the founding
Editor-In-Chief of the IEEE Transactions on Cognitive Communications and
Networking. He has also been an Editor for several journals and a member of
the Organizing or the Technical Program Committee for many international
conferences, as well as guest editor for special issues in IEEE Personal
Communications, IEEE Wireless Communications, IEEE Network and IEEE
Journal on Selected Areas in Communications. He served as a Member-atLarge of the Board of Governors of the IEEE Communications Society from
2009 to 2011, and is currently its Director of Education and Training.