Internet of Things: Architectures, Protocols, and Applications

TLDR: This survey paper proposes a novel taxonomy for IoT technologies, highlights some of the most important technologies, and profiles some applications that have the potential to make a striking difference in human life, especially for the differently abled and the elderly.

Abstract: The Internet of Things (IoT) is defined as a paradigm in which objects equipped with sensors, actuators, and processors communicate with each other to serve a meaningful purpose. In this paper, we survey state-of-the-art methods, protocols, and applications in this new emerging area. This survey paper proposes a novel taxonomy for IoT technologies, highlights some of the most important technologies, and profiles some applications that have the potential to make a striking difference in human life, especially for the differently abled and the elderly. As compared to similar survey papers in the area, this paper is far more comprehensive in its coverage and exhaustively covers most major technologies spanning from sensors to applications.

Full text

Review Article
Internet of Things: Architectures, Protocols, and Applications
Pallavi Sethi and Smruti R. Sarangi
Department of Computer Science, IIT Delhi, New Delhi, India
Correspondence should be addressed to Smruti R. Sarangi; srsarangi@cse.iitd.ac.in
Received  August ; Accepted  December ; Published  January 
Academic Editor: Rajesh Khanna
Copyright ©  Pallavi Sethi and Smruti R. Sarangi. is is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
e Internet of ings (IoT) is dened as a paradigm in w hich objects equipped with sensors, actuators, and processors
communicate with each other to serve a meaningful purpose. In this paper, we survey state-of-the-art methods, protocols, and
applications in this new emerging area. is survey paper proposes a novel taxonomy for IoT technologies, highlights some of
the most important technologies, and proles some applications that have the potential to make a striking dierence in human
life, especially for the dierently abled and the elderly. As compared to similar survey papers in the area, this pap er is far more
comprehensive in its coverage and exhaustively covers most major technologies spanning from sensors to applications.
1. Introduction
Today the Internet has become ubiquitous, has touched
almost every corner of the globe, and is aecting human life in
unimaginable ways. However, the journey is far from over. We
are now entering an era of even more pervasive connectivity
where a very wide variety of appliances will be connected to
the web. We are entering an era of the “Internet of ings”
(abbreviated as IoT). is term has been dened by dierent
authorsinmanydierentways.Letuslookattwoofthemost
popular denitions. Vermesan et al. [] dene the Internet
of ings as simply an interact ion between the physical and
digital worlds. e digital world interacts with the physical
world using a plethora of sensors and actuators. Another
denition by Pe
˜
na-L
´
opez et al. [] denes the Internet of
ings as a paradigm in which computing and networking
capabilities are embedded in any kind of conceivable object.
We use these capabilities to query the state of the object and to
change its state if possible. In common parlance, the Internet
of ings refers to a new kind of world where almost all
the devices and appliances that we use are connected to a
network. We can use them collaboratively to achieve complex
tasks that require a high degree of intelligence.
For this intelligence and interconnection, IoT devices ar e
equipped with embedded sensors, actuators, processors, and
transceivers. IoT is not a single technology; rather it is an
agglomeration of various technologies that work together in
tandem.
Sensors and actuators are devices, which help in interact-
ing with the physical environment. e data collected by the
sensors has to be stored and processed intelligently in order to
derive useful inferences from it. Note that we broadly dene
the term sensor; a mobile phone or even a microwave oven
can count as a sensor as long as it provides inputs about its
current state (internal state + environment). An actuator is a
device that is used to eect a change in the environment such
as the temperature controller of an air conditioner.
e storage and processing of data can be done on the
edge of the network itself or in a remote server. If any prepro-
cessing of data is possible, then it is typically done at either
the sensor or some other proximate device. e processed
data is then typically sent to a remote server. e storage
and processing capabilities of an IoT object are also restricted
by the resources available, which are oen very constrained
due to limitations of size, energy, power, and computational
capability. As a result the main research challenge is to
ensure that we get the right kind of data at the desired level
of accurac y. Along with the challenges of data collection,
and handling, there are challenges in communication as
well. e communication between IoT devices is mainly
wireless because t hey are generally installed at geographically
dispersed locations. e wireless channels oen have high
Hindawi
Journal of Electrical and Computer Engineering
Volume 2017, Article ID 9324035, 25 pages
https://doi.org/10.1155/2017/9324035

 Journal of Electrical and Computer Engineering
Application
layer
Network
layer
Perception
layer
Business layer
Application layer
Processing layer
Transport layer
Perception layer
A
B
F : Architecture of IoT (A: three layers) (B: ve layers).
rates of distortion and are unreliable. In this scenario reliably
communicating data without too many retransmissions is an
important problem and thus communication technologies
are integral to the study of IoT devices.
Now, aer processing the received data, some action
needs to be taken on the basis of the derived inferences. e
nature of actions can be diverse. We can directly modify the
physica l world through actuators. Or we may do something
virtually. For example, we can send some information to other
smart things.
e process of eecting a change in the physical world
is oen dependent on its state at that point of time. is
is called context awareness. Each action is taken keeping in
consideration the context because an application can behave
dierently in dierent contexts. For example, a person may
not like messages from his oce to interrupt him when he is
on vacation.
Sensors, actuators, compute servers, and the commu-
nication network form the core infrastructure of an IoT
framework. However, there are many soware aspects that
need to be considered. First, we need a middleware that can
be used to connect and manage all of these heterogeneous
components. We need a lot of standardization to connect
many dierent devices. We shall discuss methods to exchange
information and prevailing standards in Section .
e Internet of ings nds various applications in health
care, tness, education, entertainment, social life, energy
conservation, environment monitoring, home automation,
and transport systems. We shall focus on these application
areasinSection.Weshallndthat,inalltheseapplication
areas, IoT technologies have signicantly been able to reduce
human eort and improve the quality of life.
2. Architecture of IoT
ere is no single consensus on architecture for IoT, which
is agreed universally. Dierent architectures have b een pro-
posed by dierent researchers.
2.1. ree- and Five-Layer Architectures. e most basic
architecture is a three-layer architecture [–] as shown in
Figure . It was introduced in the early stages of research in
this area. It has three layers, namely, the perception, network,
and application layers.
(i) e per ception layer is the physical layer, which has
sensors for sensing and gathering information about
the environment. It senses some physical parameters
or identies other smart objects in the environment.
(ii) e network layer is responsible for connecting to
other smart things, network devices, and servers. Its
features are also used for transmitting and processing
sensor data.
(iii) e application la y er is responsible for delivering
application specic services to the user. It denes
various applications in which the Internet of ings
can be deployed, for example, smart homes, smart
cities, and smart health.
e three-layer architecture denes the main idea of the
Internetofings,butitisnotsucientforresearchon
IoT because research oen focuses on ner aspects of the
Internet of ings. at is why, we have many more layered
architectures proposed in the literature. One is the ve-
layer architecture, which additionally includes the processing
andbusinesslayers[–].evelayersareperception,
transport, processing, application, and business layers (see
Figure ). e role of the perception and application layers
is the same as the architecture with three layers. We outline
the function of the remaining three layers.
(i) e transport layer transfers the sensor data from
the perception layer to the processing layer and vice
versa through networks such as wireless, G, LAN,
Bluetooth, RFID, and NFC.
(ii) e processing layer is also known as the middleware
layer. It stores, analyzes, and processes huge amounts
of data that comes f rom the transp ort layer. It can
manage and provide a diverse set of services to the
lower layers. It employs many technologies such as
databases, cloud computing, and big data processing
modules.

Journal of Electrical and Computer Engineering 
(iii) e business la yer manages the whole IoT system,
including applications, business and prot models,
and users’ privacy. e business layer is out of the
scope of this paper. Hence, we do not discuss it
further.
Another architecture proposed by Ning and Wang [] is
inspired by the layers of processing in the human brain. It
is inspired by the intelligence and ability of human beings
to think, feel, remember, make decisions, and react to the
physical environment. It is constituted of three parts. First is
the human brain, which is analogous to the processing and
data management unit or the data center. S econd is the spinal
cord, which is analogous to the distributed network of data
processing nodes and smart gateways. ird is the network
of nerves, which corresponds to the networking components
and sensors.
2.2. Cloud and Fog Based Architectures. Let us now discuss
two kinds of systems architectures: cloud and fog computing
(see the reference architectures in []). Note that this classi-
cation is dierent from the classication in Section ., which
was done on the basis of protocols.
Inparticular,wehavebeenslightlyvagueaboutthenature
of data generated by IoT devices, and the nature of data
processing. In some system architectures the data processing
is done in a large centralized fashion by cloud computers.
Such a cloud centric architecture keeps the cloud at the
center,applicationsaboveit,andthenetworkofsmartthings
below it []. Cloud computing is given primacy because it
provides great exibility and scalability. It oers s ervices such
as the core infrastructure, platform, soware, and storage.
Developers can provide their storage tools, soware tools,
data mining, and machine learning tools, and visualization
tools through the cloud.
Lately, there is a move towards another system archi-
tecture, namely, fog computing [–], where the sensors
and network gateways do a part of the data processing and
analytics. A fog architecture [] presents a layered approach
as shown in Figure , which inserts monitoring, prepro-
cessing, storage, and security layers between the physical
and transport layers. e monito ring layer monitors power,
resources, responses, and services. e preprocessing layer
performs ltering, processing, and analytics of sensor data.
e temporary storage layer provides storage functionalities
such as data replication, distribution, and storage. Finally, the
secur ity layer performs encryption/decryption and ensures
data integrity and privacy. Monitoring and preprocessing are
done on the edge of the network before sending data to the
cloud.
Oen the terms “fog computing” and “e dge computing”
are used interchangeably. e latter term predates the former
and is construed to be more generic. Fog computing originally
termed by Cisco refers to smart gateways and smart sensors,
whereas edge computing is slightly more penetra tive in nature.
is paradigm envisions adding smart data preprocessing
capabili ties to physical devices such as motors, pumps, or
lights. e aim is to do as much of preprocessing of data
as possible in these devices, which are termed to be at the
Storage layer
Preprocessing layer
Monitoring layer
Physical layer
Security layer
Transport layer
F : Fog architecture of a smart IoT gateway.
edge of the network. In terms of the system architecture,
the architectural diagram is not appreciably dierent from
Figure . As a result, we do not describe edge computing
separately.
Finally, the distinction between protocol architectures
and system architectures is not very crisp. Oen the pro tocols
and the system are codesigned. We shall use the generic -
layer IoT protocol stack (architectural diagram presented in
Figure ) for both the fog and cloud architectures.
2.3. Social IoT. Letusnowdiscussanewparadigm:socialIoT
(SIoT). Here, we consider social relationships between objects
the same way as humans form social relationships (see []).
HerearethethreemainfacetsofanSIoTsystem:
(i) e SIoT is navigable. We can start with one de vice
and navigate through all the devices that are con-
nectedtoit.Itiseasytodiscovernewdevicesand
services using such a so cial network of IoT devices.
(ii) A need of trustworthiness (strength of the relation-
ship) is present between devices (similar to friends on
Facebook).
(iii) We can us e models similar to studying human social
networks to also study the social networks of IoT
devices.
2.3.1. Basic Components. In a typical social IoT setting , we
treat the devices and services as bots where they can set up
relationships between them and modify them over time. is
will allow us to seamlessly let the devices cooperate among
each other and achieve a complex task.
To make such a model work, we need to have many
interoperating components. Let us look at some of the major
components in such a system.
() ID: we need a unique method of object identica-
tion. An ID can be assigned to an object based on
traditional parameters such as the MAC ID, IPv
ID, a universal product code, or some other custom
method.

 Journal of Electrical and Computer Engineering
() M etainformation: along with an ID, we need some
metainformation about the device that describes its
form and operation. is is required to establish
appropriate relationships with the device and also
appropriately place it in the univers e of IoT devices.
() Security controls: this is similar to “friend list” set-
tings on Facebook. An owner of a device might place
restrictions on the kinds of devices that can connect
to it. ese are typically referred to as owner controls.
() Service discovery: such kind of a system is like
a service cloud, where we need to have dedicated
directories that store details of devices providing
certain kinds of services. It becomes very important
to keep these directories up to date such that devices
can learn about other devices.
() Relationship management: this module manages rela-
tionships w ith other devices. It also stores the types
of devices that a given device should try to connect
with based on the type of services provided. For
example, it makes sense for a light controller to make
a relationship with a light sensor.
() Service composition: this module takes the social IoT
model to a new level. e ultimat e goal of having such
a system is to provide better integrated services to
users. For example, if a person has a power sensor
with her air conditioner and this device establishes
arelationshipwithananalyticsengine,thenitis
possible for the ensemble to yield a lot of data about
the usage patterns of the air conditioner. If the social
model is more expansive, and there are man y more
devices, then it is possible to compare the data with
the usage patterns of other users and come up with
even more meaningful data. For example, users can
be told that they are the largest energy consumers in
their community or among their Facebook friends.
2.3.2. Representative Architecture. Most architectures pro-
posedfortheSIoThaveaserversidearchitectureaswell.
e server connects to all the interconnected components,
aggregates (composes) the services, and acts as a single point
of service for users.
e server side architecture typically has three layers. e
rst is the base layer that contains a database that stores details
of all the devices, their attributes, metainformation, and t heir
relationships. e s econd layer (Component layer) contains
code to interact with the devices, query their status, and use
asubsetofthemtoeectaservice.etopmostlayeristhe
application layer, which provides services to the users.
On the device (object) side, we broadly have two layers.
e rst is the object layer, w hich allows a device to connect to
other devices, talk t o them (via standardized protocols), and
exchange information. e object layer passes information to
the social layer. e social layer manages the execution of
users’ applications, executes queries, and interacts with the
application layer on the server.
3. Taxonomy
Let us now propose taxonomy for research in IoT tech-
nologies (see Figure ). Our taxonomy is based on the
architectural elements of IoT as presented in Section .
e rst architectural component of IoT is the perception
layer. It collects data using sensors, which are the most
important drivers of the Internet of ings []. ere are
various types of sensors used in diverse IoT applications.
e most generic sensor available today is the smartphone.
e smartphone itself has many types of sensors embedded
in it [] such as the location sensor (GPS), movement
sensors (accelerometer, gyroscope), camera, light sensor,
microphone, proximity sensor, and magnetometer. ese are
being heavily used in dierent IoT applications. Many other
types of sensors are beginning to be used such as sensors for
measuring temperature, pressure, humidity, medical param-
eters of the body, chemical and b iochemical substances, and
neural signals. A class of sensors that stand out is infrared
sensor s that predate smartphones. ey are now being used
widely in many IoT applicati ons: IR cameras, motion detec-
tors,measuringthedistancetonearbyobjects,presenceof
smokeandgases,andasmoisturesensors.Weshalldiscuss
the dierent types of sensors used in IoT applications in
Section .
Subsequently, we shall discuss related work in data pre-
processing. Such applications (also known as fog comput-
ing applications) mainly lter and summarize data before
sending it on the network. Such units typically have a little
amount of temporary storage, a small processing unit, and
some security features.
e next architectural component that we shall discuss
is communication. We shall discuss related work (in Sec-
tion ) on dierent communication technologies used for
the Internet of ings. Dierent entities communicate over
the network [–] using a diverse set of protocols and
standards. e most common communication technologies
forshortrangelowpowercommunicationprotocolsare
RFID (Radio Frequency Identication) and NFC (Near Field
Communication). For the medium range, they are Bluetooth,
Zigbee, and W iFi. Communication in the IoT world requires
special networking protocols and mechanisms. erefore,
new mechanisms and protocols have been pr oposed and
implemented for each layer of the networking stack, accord-
ing to the requirements impos ed by IoT devices.
Weshallsubsequentlylookattwokindsofsowarecom-
ponents: middleware and applications. e middleware cre-
ates an abstraction for the programmer such that the details
of the hardware can be hidden. is enhances interoperabi lity
ofsmartthingsandmakesiteasytooerdierentkindsof
services []. ere are many commercial and open source
oerings for providing middleware services to IoT devices.
Some examples are OpenIoT [], MiddleWhere [], Hydra
[], FiWare [], and Oracle Fusion Middleware. Finally, we
discuss the applications of IoT in Section . We primarily
focus on home automation, ambient assisted living, health
and tness, smart vehicular systems, smart cities, smart
environments, smart grids, social life, and entertainment.

Journal of Electrical and Computer Engineering 
Health and tness
Home automation
Social life &
entertainment
Applications
Near eld
communication
Wireless sensor
networks
Communication
(networking)
Middleware
Medical sensors
Mobile phone
sensors
Perception
(sensors)
Smart transport
Smart environment
Preprocessing
Chemical/
biosensors
Internet Protocol
for smart objects
Smart
agriculture
Low power
Link layer
Adaptation layer
Routing Protocol
Application Protocol
Energy
conservation
Supply chain
and logistics
Infrared sensors
Environmental
sensors
Location sensor
Movement sensors
Camera
Microphone
Light sensor
Proximity sensor
Magnetometer
Neural sensors
Service oriented
Event based
Semantic based
Database oriented
Application specic
Low power WiFi
Zigbee
Bluetooth low
energy
Low power
technologies
RFID and WSN
integration
RFID
F : Taxonomy of research in IoT technologies.
4. Related Survey Papers
Our taxonomy describes the technologies in the IoT domain
andisclassiedonthebasisofarchitecturallayers.We
have tried to cover all subareas and recent technologies in
our taxonomy. ere have been many survey papers on the
Internet of ings in the past. Table  shows how our survey
is dierent from other highly cited surveys in the literature.
Let us rst consider our novel contributions. Our paper
looks at each and every layer in the IoT stack, and as a
result the presentation is also far more balanced. A novel
addition in our survey is that we have discussed dierent IoT
architectures. is has not been discussed in prior surveys on
theInternetofings.earchitecturesectionalsoconsiders
newer paradigms such as fog computing, which have also
hither to not been considered. Moreover, our sur vey nicely
categorizes technologies based on the architectural layer
that they belong to. We have also thoroughly categorized
the network la yer and tried to consolidate almost all the
technologiesthatareusedinIoTsystems.Suchkindofa
thorough categorization and presentation of technologies is
novel to the best of our knowledge.
Alongwiththesenovelcontributionsoursurveyisfar
more comprehensive, detailed, and exhaustive as compared
to other surveys in the area. Most of the other surveys look
at only one or two types of sensors, whereas we describe
 types of sensors with many examples. Other surveys
are also fairly restricted when they discuss communication
technologies and applications. We have discussed many types
of middleware technologies as well. Prior works have not
given middleware technologies this level of attention. We
cover  communication technologies in detail and consider
a large variety of applications encompassing smart homes,
health care, logistics, transport, agriculture, environment,
smart cities, and green energy. No other survey in this area
proles so many technologies, applications, and use cases.
5. Sensors and Actuators
All IoT applications need to have one or more sensors to
collect data from the environment. Sensors are essential
components of smart objects. One of the most important
aspects of the Internet of ings is context awa reness,which
is not possible without sensor technology. IoT sensors are
mostly small in size, have low cost, and consume less power.
eyareconstrainedbyfactorssuchasbatterycapacityand
ease of deployment. Schmidt and Van Laerhoven [] provide
an overview of various types of sensors used for building
smart applications.
5.1. Mobile Phone Based Sensors. First of all, let us look at
themobilephone,whichisubiquitousandhasmanytypes
of sensors embedded in it. In specic, the smartphone is
a very handy and user friendly device that has a host of
built in communication and data processing features. With
the increasing popularity of smartphones among people,
researchers are showing interest in building smart IoT solu-
tions using smartphones because of the embedded sensors
[, ]. Some additional sensors can also be used depending