You are on page 1of 5

Bonfring International Journal of Software Engineering and Soft Computing, Vol. 9, No.

2, April 2019 31

Hybridization Approach to Classify Big Data Using


Social Internet of Things
S. Devaarul and M. Iyapparaja

Abstract--- Recently there has been quite a number of The Internet of Things (IoT) integrates a large number of
independent research activities that investigated the technologies and envisions a variety of things or objects
potentialities of integrating social networking concepts into around us that, through unique addressing schemes and
Internet of Things solutions. The resulting paradigm, named standard communication protocols, are able to interact with
Social Internet of Things has the potential to support novel each other’s and cooperate with their neighbors to reach
applications and networking services for the IoT in more common goals.
effective and efficient ways. In this context, the main Recently the idea that the convergence of the ‘‘Internet of
contributions of this paper are the following we identify Things’’ and the ‘‘Social Networks’’ worlds is possible, or
appropriate policies for the establishment and the
even advisable, is gaining momentum, as it will be discussed
management of social relationships between objects in such a in the following. This is due to the growing aware- ness that a
way that the resulting social network is navigable we describe ‘‘Social Internet of Things’’ (SIoT) paradigm would carry
a possible architecture for the that includes the functionalities
many desirable implications into a future world populated by
required to integrate things into a social network we analyze intelligent objects permeating the everyday life of human
the characteristics of the SIoT net- work structure by means of beings.
simulations.
In fact, applying the social networking principles to the
IoT can lead to several advantages the SIoT structure can be
I. INTRODUCTION shaped as required to guarantee the network navigability, so as
HERE are scientific evidences that a large number of that the discovery of objects and services is performed
T individuals tied in a social network can provide far more
accurate answers to complex problems than a single individual
effectively

(or a small group of – even knowledgeable individuals). The II. STATE OF THE ART
exploitation of such a principle has been widely investigated The collective intelligence emerging in social networks is
in Internet-related researches. As a consequence, several an extremely interesting phenomenon that has been de-
schemes have been proposed that use social networks to scribed in fascinating ways This has been pointed out by many
search Internet resources, to route traffic, or to select effective researchers – quoting the words by the DARPA Director
policies for content distribution The Internet of Things (IoT) Regina Dugan during her keynote speech at IEEE Globecom
integrates a large number of technologies and envisions a 2010 – as the key factor of a ‘‘new era of wonder for
variety of things or objects around us that, through unique science’’. In fact, the incredible success of social networking
addressing schemes and standard communication protocols, websites, such as Twitter and Facebook, and the availability of
are able to interact with each other’s and cooperate with their data about the structure and dynamics of social networks
neighbors to reach common goals Recently the idea that the collected through these websites, have attracted the attention
convergence of the ‘‘Internet of Things’’ and the ‘‘Social of a large number of scientists from several areas. In the
Networks’’ worlds is possible, or even advisable, is gaining context of communication and networking, for example,
momentum, as it will be dis- cussed in the following. This is schemes have been proposed that exploit the similarity in the
due to the growing aware- ness that a ‘‘Social Internet of interests of friends – the so called homophile – to enhance the
Things’’ (SIoT) paradigm would carry many desirable Inter
implications into a future world populated by intelligent
A first idea of socialization between objects has been
objects permeating the everyday life of human beings.
introduced by Holmquist et al. In that paper, the focus was on
In fact, applying the social networking principles to the IoT solutions that enable smart wireless devices, mostly wireless
can lead to several advantages the SIoT structure can be sensors, to establish temporary relation- ships. The authors
shaped as required to guarantee the network navigability, so as also analyze how the owners of the sensor nodes should
that the discovery of objects and services is performed control such a process. However, that work is dated 2001 and
effectively both the concepts of the IoT and the online social networks
were in their infancy.

S. Devaarul, Research Scholar, SITE School, VIT, Vellore.


More recent literature reports several researches and
E-mail: devadevvit@gmail.com experimental applications based on a new generation of
M. Iyapparaja, Associate Professor, SITE School, VIT, Vellore. objects. These enter into humans’ daily activities with a new
E-mail: Iyapparaja.m@vit.ac.in attitude and a greater awareness of the fact that they are
DOI:10.9756/BIJSESC.9018

ISSN 2277-5099 | © 2019 Bonfring


Bonfring International Journal of Software Engineering and Soft Computing, Vol. 9, No. 2, April 2019 32

designed as ‘‘smart’’ objects with a potential for inter- action We claim that these patterns of interaction among human
with each other previously inconceivable. beings are directly applicable to possible social behaviors of
typical objects that implement pervasive applications. There is
In the ‘‘things’’ connected to the Internet are clearly
distinguished from the ‘‘things’’ participating within the no doubt that many applications and services should in the
Internet of social networks, which are named with the future be associated with groups of objects whose
individuality will be ‘‘sacrificed’’ to the overall interest of
neologism Blogject, that is, ‘‘objects that blog’’. The
theoretical concept of Embodied Microblogging (EM), providing services to users (as it is the case, for example, of
introduced in, also challenges the current vision of IoTs. applications involving the use of swarm intelligence and
swarm robotics). It is equally true that many applications will
Rather than focusing on thing-to-thing or human- to-thing
interactions, it proposes two novel roles that the augmented involve an interaction among objects that will be performed
everyday objects will play: (i) mediate the human-to-human ‘‘au pair’’, i.e., where each object will be the bearer of its
specific service to the com- munity.
communication and (ii) support additional ways for making
noticeable and noticing activities in everyday life. Also the In addition, several services are already available, which
authors of show how to empower physical objects to share involve the use of multiple objects that establish asymmetric
pictures, comments, and sensor data via social networks. They relations (as, for example, in services based on Bluetooth,
also discuss about the implications of the so called ‘‘socio- Zigbee, 6LoWPAN networks of sensors/actuators or RFID
technical networks’’ in the context of the IoT. identification systems).
Last but not least, the work in introduces the idea of In other services, the objects condition their relationship of
objects able to participate in conversations that were ‘‘friendship’’ to the achievement of mutual benefits (this is the
previously reserved to humans only. Those envisioned are case, for example, of cooperative services designed to reduce
objects aware of dynamic community structures; thus, they are the energy consumption of wireless devices). Those described
able to develop a spontaneous networking infra- structure above are merely examples of services that will surely find a
based on the information to be disseminated other than the placement in the future social network of smart objects and
information on the objects themselves. that rely on the same cited relational structures that Fiske has
Recently, the idea that the IoT and the social networks are theorized for human beings.
two worlds not really that far apart from each other as one Please note that the establishment and management of such
might think, has begun to appear in the literature. It is the case relationships should occur without human intervention. This is
of the papers, for example. More specifically, in the authors not in contrast with a future vision of a ‘‘fully networked
envision the future of the Internet as being characterized by human’’. This latter is responsible only to set the rules of the
what they name Ubiquitous IoT architecture, which resembles objects, social interactions and then enjoys the services
the social organization framework (SOF) model. That work resulting from such interactions. This is a clear paradigm shift
provides an insightful overview of the expected IoT network from other proposals, according to which the objects/devices
structure. However, it does not aim at exploiting the just participate in the human social network built by their
characteristics of the social networks into the IoT. owners.
Analogously, the re- search activities reported in consider that,
By following an approach inspired by human ‘‘social
being things involved into the network together with people,
relationships’’ and ‘‘relational models’’ things mimic the
human behavior just to effectively interact with each other. A
III. A SOCIAL INTERNET OF THINGS clear advantage lies in the fact that, in so doing, models and
The cited literature, however, still lacks in some basic principles, which already proved to be effective for the study
aspects which should be addressed to fully achieve an actual of the human social networks, can be ex- tended to the object
‘‘social networks of intelligent objects’’. In fact, in analogy communities.
with the social networks of human beings we need: (i) the
Accordingly, the results shown in the last section of this
definition of a notion of social relationship among objects, (ii)
paper demonstrate that it is possible to create social networks
the design of a reference architectural model implementing a
of objects that are easily navigable, like the ones created by
Social Internet of Things based on the codified inter-object
humans.
relationships, and (iii) the analysis of the social network
structure, which derives from the objects interactions based on In Section 6 of this paper we analyze the network structure
the defined social relationships. Only a thorough investigation arising from the above types of social relationships through
of these three issues will allow for effectively extending the numerical examples.
use of models designed to study social networks of human to
social networks of things. IV. THE SIOT SYSTEM
The definition of a kind of social behavior of objects has In this section we provide an overview of a possible
been addressed in our previous works. More specifically, the implementation of the SIoT. More specifically, in Section 4.1
definition of the novel paradigm of Social Internet of Things the envisioned reference architectural model is de- scribed, in
(SIoT) and the initial studies on the relevant social structures Section 4.2 the major functions required to run the SIoT are
have been the focus of our initial investigations illustrated, and in Section 4.3 the advantages and
disadvantages of the proposed architecture are analyzed

ISSN 2277-5099 | © 2019 Bonfring


Bonfring International Journal of Software Engineering and Soft Computing, Vol. 9, No. 2, April 2019 33

4.1. The Architecture are used as an interface that rep- resents the IoT resources (i.e.
the physical world devices) and provide an access to the
To describe the proposed system we resort on the simple
functions and capabilities of these resources. Also in, an
three-layer architectural model for IoT presented . It consists
ontology is considered as a fundamental attribute of the IoT
of: (i) the sensing layer, which is devoted to the data
with the role of supporting the agent (man or machine) who
acquisition and node collaboration in short- range and local
reads an electronic tag to understand the information in it.
networks; (ii) the network layer, which is aimed at transferring
Ontologies to manage and control heterogeneous systems have
data across different networks; and (iii)the application layer,
been
where the IoT applications are deployed together with the
middleware functionalities.shows the resulting three-layer 4.1.2. Gateway and Objects
architecture. The three basic elements of the proposed system As to the gateway and objects systems, the combination of
are: the SIoT server, the gateway, and the object layers may vary mainly depending on the device
4.1.1. SIoT Server characteristics. The following three scenarios can before- seen.
In a simple one, a dummy object (e.g., a RFID tag or a
The SIoT server does not encompass the sensing layer but presence sensing device) that is equipped with a functionality
only the Network and the Application Layers. The Application of the lowest layer, is only enabled to send simple signals to
Layer consists of three sub layers. The Base Sub layer another element (the gateway). The For what concerns the
includes the database for the storage and the management of entrance of a new object into the system, the relevant activities
the data and the relevant descriptors. These record the social are mostly carried out by the object owner, who communicates
member profiles and their relation- ships, as well as the with the servers to create the account, insert the object profile
activities carried out by the objects in the real and virtual data, and set the control parameters though the ID
worlds. Data about humans (object owners as well as visitors) management and object profiling components. The ID scheme
are also managed The relevant ontologies are stored in a should be interoperable with the main identification schemes
separate data- base and used to represent a semantic view of already in use in this area, such as: IPv6 addresses, Universal
the social activities. Such a view is extracted through Product Code (UPC), Electronic Product Code (EPC), and
appropriate semantic engines. Indeed, ontology and semantic Ubiquitous code (Ucode), OpenID, URI. The profiling adds
services are necessary to provide a machine interpretable relevant information about the capabilities and history of the
frame- work for representing functional and non-functional object to its relevant ID. Given the heterogeneity of the IoT
attributes and operations of the IoT devices. In this context, nodes, SIoT members are organized in classes. Each class is
several works have been already conducted, which could be a defined on the basis of the main object features.
starting point for the definition of an ontology to be used in the
SIoT system. One solution is to adopt the Ontology Web 4.2. Main SIoT Processes
Language for Services (OWL-S) model that pro- vides both The main components of the proposed architecture are
rich expressive descriptions and well-defined semantics. This located in the Component Sub-layer. In fact, the SIoT is not
has already been used as the basis of a semantic service intended as a solution for the sensing and networking in IoT,
modeling framework for the IoT. In this framework, services but to make the world of trillions of things manageable when

ISSN 2277-5099 | © 2019 Bonfring


Bonfring International Journal of Software Engineering and Soft Computing, Vol. 9, No. 2, April 2019 34

facing the problem of service and information discovery. tasks involved in the analyzed activity (e.g., ac- count
Additionally, it aims at laying the ground for autonomous creation, profiling, parental control). These have a label
interactions among objects (mainly through service discovery associated i2j that identifies the two elements that
and composition) for the benefit of the human user. communicate to carry out the task (i; j ¼ H; S; A; O, which
In order to describe the interactions among the SIoT stand for human, SIoT server, object agent, and object,
respectively). Notice that herein the gateway is not mentioned,
architectural elements, in we provide an overview of the
processes related to four main SIoT activities, namely: even if it may take part in these processes when the agent is
entrance of a new object, service discovery and composition, involved. This is because, in this context, the agent is defined
as the software entity that implements the application
new object relationship establishment, and service
provisioning. In the figure, the square blocks represent the functionalities of either the object or the gateway.

4.3. Analysis of the Proposed SIoT System and social relationships can be detected.
The IoT domain is characterized by a significant frag- The discovery process in our approach is driven by the
mentation and by the presence of heterogeneous systems relationship links among objects, which are followed to find
based on dissimilar architectures. This makes a synergistic the target service providers. Once these are found, their
integration process difficult to be carried out. The need for a trustworthiness levels are evaluated to select the most reliable
clear reference architectural model that will allow the different ones. The amount of interaction between the service discovery
systems to cooperate is, thus, strongly felt. As such a model is and the relationship management modules is limited; for this
still missing, we tried to adhere to the following principles: reason, we have decided to keep these two modules separate.
define an architecture that could foster the interoperability However, if the discovery is performed by following the links
with existing IoT components, proto- cols, interfaces, and involving trusted nodes only (i.e., there is a need for
functionalities; include mechanisms for the efficient navigation across ‘‘trusted’’ areas), then the number of
integration of this architecture into the service layer of the interactions between the two above components may increase
Future Internet networking infrastructure. significantly. This may reduce the efficiency in resource
The relationship management functionality in our solution discovery.
is implemented only into the Server, without a collaboration of
the gateway and the objects. We decided for this solution to V. SAMPLE APPLICATIONS
allow ‘‘non SIoT-enabled’’ devices and relevant gateways to Several applications can benefit from the availability of
take part in the SIoT without the need for updating their social relationships between things interconnected to a
systems. This approach, however, has the disadvantage of network composed of trillions of nodes. While a few
requiring a continuous communication with the servers for the interesting applications can be already defined, many others
creation and the update of the relationships. In particular, it is will show up in the years to come according to the increase in
necessary to send information about the activity of the owner the number and categories of objects able to connect to the
monitored by the objects (e.g., position, use of the Inter- net Internet. Provides a sketch of some sample applications, which
connection, movements), so that the co-location, co- work, are described in the following:

ISSN 2277-5099 | © 2019 Bonfring


Bonfring International Journal of Software Engineering and Soft Computing, Vol. 9, No. 2, April 2019 35

[5] M. Iyapparaja and M. Tiwari, “Security policy speculation of user


VI. CONCLUSIONS uploaded images on content sharing sites”, IOP Conf. Series: Materials
Science and Engineering, 2017.
In this paper we have focused on the integration of social [6] M. Iyapparaja and P. Sivakumar, “Metrics Based Evaluation for Disease
networking concepts into the Internet of Things, which leads Affection in Distinct Cities”, Research J. Pharm. and Tech., 2017.
to the so called ‘‘Social Internet of Things’’ (SIoT) paradigm. [7] M. Iyapparaja, “Coupling and Cohesion Metrics in Java for Adaptive
Reusability Risk Reduction IET Chennai”, 3rd International Conference
Recently, the SIoT has been the subject of several independent on Sustainable Energy and Intelligent Systems, Pp. 52-57, 2012.
research activities as it promises to achieve scalable solutions [8] M. Srivathsan and K.Y. Arjun, “Health monitoring system by prognotive
in networks interconnecting trillions of nodes and to support computing using big data analytics”, Procedia Computer Science, Pp.
new interesting applications. More specifically, in this paper 602-609, 2015.
[9] P. Tague and R. Poovendran, “A canonical seed assignment model for
we have identified the types and the characteristics of the key predistribution in wireless sensor networks”, ACM Transactions on
social relationships that can be established by objects in the Sensor Networks, Vol. 3, No. 4, 2007.
SIoT. Furthermore, we have proposed a system architecture [10] P. Tague and R. Poovendran, “Modeling adaptive node capture attacks
and the required basic functionality for an implementation of in multi-hop wireless networks”, Ad Hoc Networks, Vol. 5, No. 6, Pp.
801-814, 2007.
the SIoT. Finally, we have statistically analyzed the structure [11] W. Feller, “An Introduction to Probability Theory and Its Applications”,
of the SIoT network. Our analysis has been based on the John Wiley & Sons, Inc., 1957.
output of the SWIM mobility simulator. Results of such an
analysis show that the probability distributions of the distance
between nodes that are linked by a social relationship depend
on the type of relationship. More specifically, For OOR
relationships, such a distribution is characterized by a power
law behavior.

REFERENCES
[1] B. Kayyali, D. Knott and S. Van Kuiken, “The big-data revolution in US
health care: Accelerating value and innovation”, Mc Kinsey &
Company, Vol. 2, No. 8, Pp. 1-13, 2013.
[2] H.A. David and H.N. Nagaraja, “Order Statistics”, 3rd Ed. New Jersey:
John Wiley & Sons, Inc., 2003.
[3] J. Archenna and E.A. Marry anita, “A Survey of Big Data Analytics in
Healthcare and Governments”, Science Direct, 2015
[4] M. Iyapparaja and B. Sharma, “Augmenting SCA project management
and automation Framework”, IOP Conf. Series: Materials Science and
Engineering, 2017.

ISSN 2277-5099 | © 2019 Bonfring