Hindawi

Wireless Communications and Mobile Computing

Volume 2018, Article ID 8739203, 16 pages
https://doi.org/10.1155/2018/8739203

Research Article
Clustering Approaches for Pragmatic
Two-Layer IoT Architecture

J. Sathish Kumar and Mukesh A. Zaveri

Computer Engineering Department, SVNIT, Surat, India

Correspondence should be addressed to J. Sathish Kumar; [email protected]

Received 13 September 2017; Revised 8 February 2018; Accepted 11 March 2018; Published 19 April 2018

Academic Editor: Mauro Femminella

Copyright © 2018 J. Sathish Kumar and Mukesh A. Zaveri. This 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.

Connecting all devices through Internet is now practical due to Internet of Things. IoT assures numerous applications in everyday
life of common people, government bodies, business, and society as a whole. Collaboration among the devices in IoT to bring
various applications in the real world is a challenging task. In this context, we introduce an application-based two-layer architectural
framework for IoT which consists of sensing layer and IoT layer. For any real-time application, sensing devices play an important
role. Both these layers are required for accomplishing IoT-based applications. The success of any IoT-based application relies on
efficient communication and utilization of the devices and data acquired by the devices at both layers. The grouping of these
devices helps to achieve the same, which leads to formation of cluster of devices at various levels. The clustering helps not only
in collaboration but also in prolonging overall network lifetime. In this paper, we propose two clustering algorithms based on
heuristic and graph, respectively. The proposed clustering approaches are evaluated on IoT platform using standard parameters
and compared with different approaches reported in literature.

1. Introduction The most significant part of above-mentioned applica-

tions using IoT needs sensing and monitoring the envi-
The success of wireless sensor network in the form of technol- ronment or acquiring the data from different IP-enabled
ogy and applications in different areas like home automation, devices or sensors. The sensor devices used for sensing and
industrial applications, security and military surveillance, monitoring are battery-operated and energy-constrained.
and many more raises the further need for machine-to- This implies that power consumption and energy are critical
machine connectivity and availability of the data or infor- aspects. IP-based communication effectively utilizes more
mation anytime and anywhere [1]. This requirement leads energy, which leads these low-powered devices to deplete
to the new technology development in the form of Internet rapidly. These huge numbers of devices communicate and
of Things (IoT). The IoT allows the connectivity among the collaborate with each other in order to accomplish a given
devices and helps in acquisition of data or information at any job or task. This raises the need for effective connectivity
time and from any place. There are numerous applications and efficient communication among these devices in an
of IoT coming up using different technologies [2]. Also, optimized way, which is a very challenging task. In this
IoT enables innovative services in numerous applications context, there is a need for a solution that promises maximal
like smart transportation, smart home, smart city, smart connectivity through minimum communication. The most
lifestyle, smart retail, smart agriculture, smart industries, efficient way to fulfill these needs is to collaborate among
smart emergency, smart health care, smart environment, and the devices or sensors and perform the tasks for a given
many more [3, 4]. The use of these applications and their application. One way to achieve this is through grouping
demand has increased the scope of research and innovation the devices in an efficient way in terms of energy usage and
in this domain [2, 5]. computational complexity.
2 Wireless Communications and Mobile Computing

For any IoT-based application, it is required to collect the As pointed out earlier, the nodes are tiny and energy-
data through sensing devices and process these data through constrained. In this context, it is desirable to form a group
different algorithms. Then, the processed information can of devices and the group of devices communicate among
be accessed through the Internet anywhere and at any time. themselves through a group leader that ensures the minimum
The grouping of devices or sensors is known as clustering. communication overhead and energy-efficient execution
Clustering helps in carrying out the task of acquiring the using clustering. In the literature, clustering in wireless sensor
information in an efficient way with minimum number of network (WSN) has been explored very well [11–25]. In our
communications within a network [6] and disseminating this paper, we also address the clustering of sensor nodes but in
information for further processing. Clustering also helps in the Internet of Things environment. There are very important
prolonging the network lifetime and further the lifetime of facts to be noted which differentiate the clustering in WSN
an IoT-based application that is deployed for a specific task. compared to that in IoT. The clustering in WSN is performed
For efficient clustering in the IoT environment, there is a locally in the network, whereas clustering in IoT may be
need to have (i) classification between the underlying sensor considered as globally in the network. The number of param-
devices and IP-enabled devices and (ii) minimum commu- eters restricted to WSN may be connectivity and density of
nication overhead for accessing the data. To achieve this, the nodes, which are either layer-based [26, 27] or without
we introduce two-layer architectural framework consisting layer and hence very much constrained by the mobility of the
of two layers: upper layer that consists of IP-enabled IoT nodes. But in the case of IoT environment, the density is not
devices and lower layer that is simply sensor devices. In this a constraint. Moreover, the mobility of IoT nodes allows on
regard, for grouping the devices, clustering algorithms are the fly clustering with underlying sensor nodes dynamically
employed. In clustering, each cluster chooses one node within in the region of interest, which is not possible in case of
the cluster as a cluster head, and further the communication WSN. Further, in the literature, there is no method reported
among the nodes in a network is carried out through the for clustering in IoT environment so far, and, for the first
cluster head. If two nodes in a network are not within the time, it is explored using the two-layer architecture in IoT as
range of each other, multihop communication is required. introduced in our earlier proposed work [9, 10].
Communication through cluster head avoids multihop com- In [26], cluster-based architecture has been proposed to
munication up to some extent. The cluster head also helps in structure the topology of different wireless sensor networks
aggregating the data acquired by different nodes in a network. to coexist in the same environment by creation of virtual
In IoT-based network, as mentioned, two-layer architecture wireless sensor networks. Similarly, layer-based clustering
consists of sensor layer and IoT layer. The grouping of these approach has been reported in [27] for routing with homo-
devices is possible in two ways, that is, from sensor layer geneous and densely deployed network using cross-layer
to IoT layer and vice versa. For accommodating both these interaction. The approach in [26] does not support the
approaches, we propose two clustering algorithms, namely, discovery of nodes on the fly which is mandatory for the
heuristic-based and graph-based clustering, in this paper, network with mobility to enable anytime and anywhere data
respectively. accessibility, which can be supported by IoT-based network.
The rest of the paper is organized as follows. The literature In our proposed method, there is no restriction of dense
survey with respect to related work regarding the different network deployment compared to the method in [27], and,
clustering techniques is described in Section 2. Section 3 moreover, our proposed approach also supports clustering in
introduces a layered architectural framework for IoT. This the case of heterogeneous environment.
architectural framework is a system model for our proposed The literature related to different clustering approaches in
clustering algorithms. Section 4 describes the proposed clus- the wireless domain is referred to here, which motivates us
tering approaches. The simulation results and experiments to propose a clustering method for a layered IoT framework.
are detailed in Section 5. Section 6 concludes the paper. Various clustering algorithms are available in the literature
and are proposed by considering the distributed mechanisms
2. Related Work with respect to WSN. In low-energy adaptive clustering
hierarchy (LEACH) [11], two-level-LEACH [12] and energy-
In the literature, efforts have been noticed to propose the IoT efficient hierarchical clustering (EEHC) [13], residual energy
architecture [7, 8] in the form of generic architecture or as is not considered, which leads to unstable network. Energy-
reference model. These architectures have been introduced at efficient unequal clustering (EEUC) [14], power-efficient
conceptual level without any implementation and validation. and adaptive clustering hierarchy (PEACH) [15], multihop
The generic architecture or reference model is too complex routing protocol with unequal clustering [16], and EEHC
for one-to-one mapping for any real application. Earlier, IoT [13] require 𝑘 number of hops to reach cluster head (CH)
architecture [9, 10] has been presented in an abstract way and with overhead in joining the cluster. Hybrid energy-efficient
evaluated to measure the performance. In practice, there is distributed clustering (HEED) [17] has higher overhead for
a need to integrate ordinary sensors or IoT devices with an CH election. Sensor Web (S-WEB) [18] is a hybrid technique
application. In this context, a system model using two-layer of centralized and distributed clustering, where most of the
IoT framework is introduced and the communication cost in tasks are performed by the nodes, except the beacons which
this model may be reduced to one or at the most two hops for are generated from the base station. Ding et al. [25] have pro-
any node in a network. This two-layer framework is described posed distributed weight-based energy-efficient hierarchical
later in the next section. clustering (DWEHC) approach to enhance over the HEED
Wireless Communications and Mobile Computing 3

approach by assigning weight to the parameters. Clustering information through sensing and send it to the CH. CHs
approaches incorporating the mobility have been proposed communicate the received information to IoT node. Besides,
in [19–24]. LEACH is designed for sensor network without all the cluster members are at one hop to CHs and all the
considering the mobility of the nodes. The approach, namely, CHs are at one hop to IoT node, which leads to minimum
LEACH-mobile [19, 20], has been proposed incorporating communication overhead.
the mobility of the nodes. Similarly, Nagpal et al. have also In brief, the cluster head performs receiving and for-
proposed CLUBS clustering approach [21, 22] for WSN by warding of sensed data or aggregated data. The aggregation
incorporating the mobility of the nodes. In this approach, of data reduces overall communication within a network,
the clusters are formed by using local broadcasting and local as CH communicates to base station instead of all nodes
density of nodes is used as criteria for cluster formation. In deployed in the network. Cluster members also help in
CLUBS approach [21], the cluster formation has been carried scheduling communication whenever it is essential in order
out considering three properties: (i) every node should come to save energy consumption of the network. The cluster-
under a cluster; (ii) maximum value of diameter should be based network topology is simple and easy to manage and is
equal for each cluster; and (iii) intracluster communication conducive for running an application in distributed manner.
must be allowed. Manjeshwar and Agrawal proposed an The cluster is also very helpful in a dynamic environment,
event-driven clustering approach called threshold-sensitive where a node leaves or enters the cluster and becomes part of a
energy-efficient sensor network protocol (TEEN) [28]. In cluster which is adaptive in nature in the sense that each node
this approach, the sensed data is forwarded to the base decides whether to join the cluster or become a CH. If CHs
station only if some event occurs, which is based on two and cluster formation are changed with time or rounds, then
thresholds, that is, soft and hard. The disadvantage of this it is called dynamic cluster formation, whereas in static cluster
approach is that the node responds only if the change in formation, once the clusters are created, they remain the same
the attributes crosses these threshold values, which makes throughout the network lifetime. Dynamic clustering implies
the approach less applicable in dynamic environment, as the frequent changing of the cluster formation and the respective
selection of two threshold values is very sensitive and difficult CH selection at different time intervals. Although frequent
for real applications. The user may keep on waiting to get cluster formation is carried out due to adaptive nature of
response and does not get any information about the status the network, there is also an advantage of it with respect to
of the node, which makes this approach not suitable for energy. As dynamic clustering allows selecting different CHs
the applications, where the periodic updates are required. at different time intervals, the energy level of different nodes
Later, this approach has been enhanced and proposed as can be incorporated while making such selection as battery
adaptive threshold-sensitive energy-efficient sensor network lifetime of the node is very critical for overall existence of the
(APTEEN) [29]. APTEEN combines event-driven approach network. Thus, dynamic clustering promises to balance the
of TEEN and periodic approach of LEACH to address the energy level among the different nodes across the different
problems occurring in TEEN. APTEEN is good for periodic clusters in the network.
applications, but the complexity of the approach increases To overcome the above limitations and for accommo-
due to inclusion of extra threshold function and count time. dating the various requirements for minimum communica-
Cluster head election using fuzzy logic (CHEF) has been tion and maximum collaboration among the sensing nodes
proposed by Kim et al. [30], which is an enhanced approach and IoT nodes, in this paper, we propose two clustering
over LEACH using two parameters, that is, distance and approaches that are most suitable for IoT architectural
residual energy for cluster head selection. This approach framework. These two approaches are, namely, heuristic-
does not consider the intercluster communication. Bagci and and graph-based clustering, respectively. For evaluating the
Yazici [31] proposed energy-aware unequal clustering with architectural framework and clustering methods introduced
fuzzy (EAUCF) approach, which is a distributed competitive in this paper, the proposed algorithms are compared thor-
unequal clustering algorithm. In this approach, the selection oughly with the existing methods in literature using the
of cluster head is performed based on the competition radius, standard set of parameters like power consumption, com-
which is determined by distance and energy among the munication cost, number of cluster heads, and number of
different nodes using probabilistic and fuzzy techniques. This hops. The experiments are executed in IoT-based platform
algorithm assumes that working area of each node is directly using Contiki-2.7 operating system and Cooja simulator [32].
proportional to its energy; otherwise node will die rapidly. For summarizing the contribution in this paper, we would
Many distributed approaches are also available in liter- like to emphasize that our proposed IoT-based clustering
ature, but it is not possible to use these algorithms in the is based on only two-layer architecture, that is, underlying
current form for IoT framework. The reason is that these sensor layer and IoT layer, which allows clustering globally
algorithms do not consider different parameters that are in the network in the region of interest and provides the
necessary for IoT environment like heterogeneity, support for anytime, anywhere data accessibility and on the fly dynamic
IP, and mobility. Mobility requires dynamic cluster formation. clustering due to the inclusion of mobility in the approach at
It is also important to note that IoT network consists of IP- IoT layer. There are approaches in the literature reported for
enabled devices and, in such cases, one hop communication clustering with mobility in the wireless sensor network but
is an ideal case and the most desirable one. Multiple IoT these approaches are operated locally in the network, whereas
nodes act as base stations for underlying sensor nodes, which IoT-based approach supports the global network scenario
enables one hop communication. These devices acquire the anytime and anywhere.
4 Wireless Communications and Mobile Computing

IoT server
IoT cloud

IoT
IoT layer
IoT processors and controllers Smartphones

Sensing
layer

Sensing devices deployed in the network

Figure 1: Two-layered hierarchical IoT architectural framework.

3. System Model ordinary nodes may communicate through local group leader
called cluster head (CH) to IoT device and the data acquired
In this section, the two-layer IoT architectural framework, through ordinary sensor node may now be accessible any-
which represents a system model for proposing the clustering where and anytime. For dynamic scenario, the mobility may
algorithms, is described. The sensor network is mainly used be incorporated in few IoT nodes whenever it is necessary
for acquiring the data from the surrounding environment [39]. In this view, a two-layered IoT architectural framework
depending upon the applications. IoT network consists of is introduced.
many sensing devices that are not necessarily to be connected The architectural framework for IoT applications consists
to the Internet. In this view, it is preferred to have a of two layers, namely, IoT layer and sensing layer. Sensing
framework that differentiates between IP-enabled devices layer is deployed with devices that are either IP-enabled or
and non-IP-enabled devices, that is, IoT devices and simple ID-enabled, depending on the requirement of the application.
sensing nodes without IP capability. This kind of framework In this layer, deployed devices include sensors, actuators,
provides a layered architecture and is efficient in terms of and RFID devices. An IoT layer device comprises IP-enabled
communication for exchanging the data, which is validated devices with IoT protocol stack which is significant in nature,
through simulations described later in this paper. because the IoT protocol stack [40] has been introduced to
Generally, in wireless sensor network, the nodes that are operate in energy-constrained environment at any layer in the
more than one hop away from the base station or access network hierarchy, that is, at data link using IEEE 802.15.4e
point consume energy rapidly [33–38]. In IoT framework, if [41], networking using 6LowPAN [42], routing using RPL
sensor motes or nodes are static to acquire the data which [40], or application layer using CoAP [43], which results in a
should be accessible anytime and anywhere, the multihop huge difference in developing the clustering approach in IoT
communication should be handled carefully for optimizing as compared to that in WSN.
network resources for prolonging the lifetime of an appli- These IoT devices are expected to have longer battery life
cation deployed for a specific task. It is preferred that IoT and storage with the ability to perform real-time processing
devices that have more energy and higher end processors and communication, as compared to the functionalities
as compared to underlying ordinary sensor nodes should provided by ordinary nodes. These features of IoT devices
be available to underlying sensor nodes for communication are also necessary for availability of the data or information
for at most two hops. The IoT devices may have mobility acquired anywhere and anytime. An important job of mobile
and this provides further flexibility to underlying static IoT nodes is to monitor and collect the information from CHs
nodes for communication, so that many such static nodes in the sensing layer. The two-layered IoT framework is shown
in the network may be covered by mobile IoT devices. Thus, in Figure 1.
Wireless Communications and Mobile Computing 5

Communication between the devices in the IoT layer and proposed two-layer framework, the clusters can be formed by
sensing layer consists of different possibilities. As shown in either of the two layers. In the sensing layer, the nodes form
Figure 1, one CH may communicate with one IoT node; two the cluster and determine the CH without intervention of
CHs may communicate with one IoT node; and one CH may IoT nodes, according to the typical and traditional clustering
communicate with two IoT nodes, depending on whether the mechanism. The approach is called heuristic-based clustering
IoT node is within its range of transmission or not. Although algorithm that works from bottom to top. The other way of
every node in sensing layer may not be in the range of other cluster formation and electing CH is by IoT nodes. Assuming
nodes, still all the sensing nodes are capable of understanding that the IoT node has information of sensing nodes that
the scenario of entire network through communicating and are within the range of the IoT node, it forms the cluster
collaborating with IoT nodes. Assuming that thousands of and decides optimally the CH. After cluster formation and
devices are scattered in the real-time network and are going CH selection, the sensor nodes are able to communicate
to take part in accomplishing the given task, communication with the IoT node through the CH of respective cluster.
and connectivity should be addressed in an energy-efficient This approach can be considered as top to bottom approach
way. In this regard, the proposed clustering mechanisms for as IoT layer performs the task of cluster formation and
our two-layer architecture are more suitable for any real-time CH selection. To perform this task, a graph-based clustering
IoT-based application, which is elaborated in detail in the algorithm is proposed, where clustering is carried out in
following section. the sensing layer with the involvement of IoT nodes. In
From this discussion, it is clear that the IoT network the graph-based clustering approach, IoT nodes form the
may generate a huge amount of data over the time, and thus clusters and select CH, which reduces overall communication
accessing the same anytime and anywhere is a challenging cost at sensing layer and hence reduces energy consumption,
task. In such a scenario, it is important to have an IoT-based which helps in prolonging the network lifetime. The two
cloud environment that facilitates storing and processing the proposed approaches are generic and can be applied for any
sensing information in the cloud, where IoT nodes perform application depending on the need. Further, the proposed
real-time processing to accomplish a given task and forward approaches work on layered architecture that, in turn, helps
the data to corresponding static IoT nodes or IoT cloud. IoT manage the nodes, performs tasks optimally, and reduces the
cloud is connected to IoT servers and IoT nodes are located communication overhead.
physically anywhere. IoT server is basically an IoT node that The clustering is performed based on a number of
is responsible for high level data processing that helps to make parameters. Generally, these parameters are connectivity,
appropriate decisions. Similarly, IoT smartphone is also an distance between the nodes, residual energy of a node, and
IoT node that can be used to access the acquired information so forth. For our proposed algorithm, we consider two
and to control the applications remotely. Likewise, IoT parameters, namely, the number of neighbors and residual
processors and controllers are also considered as IoT nodes energy of a node. The reason behind considering the number
for small scale card sized computers or embedded processors. of neighbors as one of the parameters is that the cluster
For simplicity, the IP-based processors, controllers, and should be formed using the neighbor nodes only. In our
vehicles may be called IoT microcomputers, IoT controllers, approach, the neighbor node is decided based on the radius
and IoT vehicles, respectively. Also, they can be used as of transmission range and the nearby nodes, whose beacon
parameters on IoT cloud and can be utilized for various satisfies a certain SNR that is more justified way to decide
applications. Based on applications, the device mobility also whether the given node is a neighbor node or not. Simply,
plays an important role in IoT environment. Few of these the connectivity-based clustering results in the set of nodes
devices may be static and others may be mobile. For instance, in the cluster which are multiple hops apart; still it forms
IoT cloud and IoT server can be considered as a static node, a cluster. This is not a good solution from energy point of
whereas IoT smartphones, IoT vehicles, IoT microcomputers, view, as multihop communication within a cluster consumes
and IoT controllers can be considered as mobile nodes. This more energy. The best way is to have one hop neighbor
implies that these IoT nodes are portable and their positions and form a cluster. The same scenario is possible using the
vary at different times. This IoT-layered framework is used for selection of neighbor nodes based on transmission range and
developing the clustering algorithms, which is detailed in the SNR as mentioned above. Moreover, the residual energy is
following section. very important parameter as the node with more residual
energy should be chosen as cluster head. In our approach,
we consider the residual energy as another parameter and
4. Proposed Approaches the selection of cluster head based on it allows one hop
communication to IoT layer, which results in energy-efficient
As discussed earlier in this paper, the devices in the network communication.
are going to perform various tasks based on application. An If a given node has more residual energy, then its
important issue is to have energy-efficient connectivity and transmission range and SNR with respect to other nodes will
communication. This can be achieved through clustering or be more and hence the chance of inclusion of such nodes in
grouping various devices in the layered IoT framework. Any the cluster is more. In this way, the residual energy is related
clustering algorithm needs to perform two steps: (i) cluster to the selection of the number of neighbor nodes. Further, the
formation and (ii) CH selection. The cluster formation and more the number of neighbor nodes in the cluster is, the more
CH selection can be carried out in two ways. Based on the the coverage of the network deployed in the region of interest
6 Wireless Communications and Mobile Computing

is. Due to lesser number of cluster heads, the computational performs the cluster formation and elects CH in a sequence of
complexity is minimized as minimum number of cluster three steps. These three steps are broadcasting, multicasting,
heads are communicating with the IoT devices in upper and unicasting, respectively. As discussed earlier, in our IoT
layer. It is important to note that this communication takes framework, there are two types of devices: non-IP-enabled
place with one hop only, which is a major advantage of our devices and IP-enabled devices. The non-IP-enabled devices
proposed hierarchical two-layer architecture. are addressed using identification number locally by assigned
The number of neighbor nodes is calculated as follows. ID and for IP-enabled devices, IP addresses are assigned using
The number of neighbors for each sensor node 𝑖 is determined IoT protocol stack [40]. Each step is described as follows:
by the following equation:
(i) In broadcasting step, every node broadcasts the
𝑘 packets or sends the beacon signals that consist of
Neighbours Count (i) = ∑ 𝑋𝑗 , (1) its ID or IP address, so that each node may be
𝑗=1 informed of its neighbors within its radio range of
transmission. By this way, all nodes come to know
where 𝑋𝑗 = 1, if node 𝑗 is within the radius of transmission about their neighbors in the network. Besides, each
range with desirable SNR and 𝑋𝑗 = 0, otherwise. node maintains the neighbor list and total neighbors
The term 𝑘 denotes the total number of nodes within can be determined by the total count using (1).
default transmission range. All nodes within the transmission
(ii) In the second step, multicasting is performed. Each
range of a node 𝑖 do not have desirable SNR and hence those
node sends the neighbor count and residual energy to
nodes that have desirable SNR are considered for neighbor
its neighbors along with their ID or IP address within
counting. The desirable SNR can be set to a particular value
its radio range of transmission as determined using
depending on the application in hand. The residual energy
(1) to (5).
is calculated using radio energy dissipation model [44] and
power loss that is proportional to (distance)2 in the free (iii) Finally, in third step, cluster formation and selec-
space and to (distance)4 in the case of multipath fading, tion of CH are carried out. Hence, after receiving
respectively. The distance is represented by 𝑑. The energy the information about their neighbors, every node
consumption for transmitting 𝑚 bits message over distance maintains the neighbors list and neighbors count. By
𝑑 can be formulated as shown in (2) and (3), respectively. determining the maximum count of the number of
neighbors and residual energy (RE), CH is elected.
𝐸Tx = 𝑚 × (𝐸elec + 𝜀fs × 𝑑2 ) , if free space (2) Consequently, the cluster is formed by the elected
cluster head using the neighbors list it has, which
𝐸Tx = 𝑚 × (𝐸elec + 𝜀amp × 𝑑4 ) , if multipath, (3) consists of all nodes within its radio range of trans-
mission. The flowchart of the proposed algorithm is
where 𝜀fs is power consumption of the free space propagation, presented in the Figure 2. The corresponding algorith-
𝜀amp is power consumption of multipath propagation, and mic steps are presented in the form of pseudocode in
𝐸elec represents the residual energy of network. To receive Algorithm 1.
𝑚 bits of information, the radio expends as shown in the
following equation: The proposed algorithm is demonstrated using an illus-
trative example with three nodes as follows. Let us assume
𝐸Rx = 𝑚 × 𝐸elec . (4) three nodes in the network, as shown in Figure 3, such that
node 1 within its range has only one neighbor, that is, node
Therefore, the residual energy for each sensor node 𝑖, RE(𝑖), is 2. Node 2 has two neighbors within its range, that is, node 1
determined in our approach by using the following equation: and node 3. Likewise, node 3 has only one neighbor, that is,
node 2. Now, every node exchanges node (NC and RE) with
RE (𝑖) = 𝐸0 (𝑖) − (𝐸Tx (𝑖) + 𝐸Rx (𝑖) + 𝐸Computation (𝑖)) , (5) all other neighbors and calculates the maximum of NC and
RE values, where NC is neighbors count and RE is residual
where 𝐸0 is initial energy of the node and 𝐸Computation is energy.
the amount of energy consumption in local processing while These maximum values are used as reference values in
executing cluster head selection algorithm on respective cluster head selection process. If nodes have maximum NC
nodes. value but less RE value is selected as cluster head, then the
network becomes unstable due to CH die-out over the time.
4.1. Heuristic-Based Clustering Approach. Heuristic-based Similarly, if nodes have maximum RE but less NC value,
clustering approach is introduced using the parameters then it may result in more number of clusters in the network
described above, namely, the neighbors count and residual and, hence, result in more communication overhead. So, it is
energy. Neighbors count also indicates connectivity of the required to consider both values simultaneously for cluster
nodes. The cluster formation is within the radio range head selection. For this, max(NC, RE) function is defined,
transmission of a node. Let us assume that the nodes are which returns the node number 𝑖 having maximum value of
deployed randomly in the network with unit disk graph NC and RE combined together.
medium (UDGM) model [45]. The network is assumed to be Node 1 determines max(Node 1 (NC = 1, RE),
dynamic; that is, the nodes may have mobility. The algorithm Node 2 (NC = 2, RE)). It finds the maximum and determines
Wireless Communications and Mobile Computing 7

Start

Deploy N nodes randomly with mobility

At regular While
interval of times periodically

Till all the nodes are For each

considered in the node i ∈ N
network

Find neighbors in the transmission range

Share neighbors count (NC) and residual

energy (RE) with all the neighbors

Choose cluster head (CH) by determining

the maximum value of NC and RE

Form a cluster within the range of CH

Update the cluster formation and CHs

CHs communicate with the IoT nodes

Stop

Figure 2: Flowchart of heuristic algorithm.

Input Data: 𝑁 //Number of nodes in sensing layer

Result: CH’s election and Cluster’s formation
(1) Deploy the nodes randomly with UDGM model;
(2) Set the node mobility parameters;
(3) while Periodically do
(4) for each node 𝑖 ∈ 𝑁 do
(5) Step 1 → Send (“random messages to find the neighbors”);
(6) Step 2 → Send (NC, RE) to all neighbors;
// Where NC = Neighbor count and RE = Residual Energy are calculated using equation (1) to (5);
(7) Step 3 → Determine CH = max(NC𝑗 , RE𝑗 ), //where 𝑗 is total number of neighbors of 𝑖
(8) Form a cluster within the range of CH.;
(9) end
(10) Update the Cluster formation and CHs;
(11) end
(12) CH’s Communicate and collaborate with the IoT nodes;

Algorithm 1: Heuristic-based clustering algorithm.

8 Wireless Communications and Mobile Computing

nodes simply reject them and the packets must be received

2 by their family members only.
1 3 In case 2, nodes need to communicate with other family
types. In this case, different varieties of nodes communicate
and collaborate across different family type of nodes. This is
a typical scenario representing heterogeneous environment.
Figure 3: Example of cluster head selection and cluster formation. Nodes must be having the same type of protocol for commu-
nication. In the context of IoT, if these nodes are IP-enabled
to communicate with other node types, then this leads us
to a nonoverlapping cluster formation. CH is selected from
CH as node 2. Likewise, node 2 computes max(Node 1 (NC = the clusters depending on the energy and connectivity. For
1, RE), Node 2 (NC = 2, RE), Node 3 (NC = 1, RE)) and instance, assume that temperature and humidity nodes are
determines node 2 as CH. Node 3 calculates max(Node 2 (NC deployed in the network. In such a situation, nodes can com-
= 2, RE), Node 3 (NC = 1, RE)) and determines CH as node municate with each other to determine weather conditions.
2. Hence, CH is selected as node 2 and cluster is formed The nodes have to communicate and process the information
within its radio range of transmission. In this example, in a collaborative fashion to make the decisions effectively.
it is assumed that all nodes have same residual energy; IP-enabled IoT devices allow communicating among the
that is, RE1 = RE2 = RE3 = RE. RE𝑖 is residual energy different family types of nodes through IoT layer, which
of node 𝑖. In real scenario, this is not possible, so our is an advantage of two-layer framework, even in the case
Max function evaluates the nodes using both NC and RE of heterogeneous network. The advantage of our proposed
values simultaneously and returns the corresponding node heuristic-based clustering algorithm is that it is simple, easy
number. This is possible because NC and RE values are to implement, and provides effective communication and
exchanged among all nodes within the transmission range of connectivity within the network as well as across the network
a particular node. Suppose that if node 2 has lesser residual through IoT layer.
energy, then it cannot be selected as cluster head, though the The heuristic-based clustering is carried out mainly by
NC is maximum compared to that of other nodes. In such the sensing layer and the IoT layer has information about
a case, instead of node 2, the node that has the maximum CH in the sensor layer. It results into overhead for sensing
residual energy may be selected as cluster head. In the above layer as both the tasks of clustering and data acquiring are
mentioned example, either node 1 or node 3 is elected as to be handled by the sensing layer only. This scenario is
cluster head and forms a cluster. In this way, the cluster head good enough for static clustering; that is, nodes are static,
is selected and the cluster formation is performed. Likewise, and once the cluster is formed, it remains unchanged. In
remaining nodes form other clusters. case of dynamic network, where nodes are mobile in the
Further, in our proposed system, we accommodate the sensing layer, the clustering process needs to be performed
formation of clusters that are based on heterogeneous and whenever the topology changes due to the movement of these
homogenous nodes. The reason is that a particular applica- nodes. So, it may result into more energy consumption in
tion may need to exchange the data acquired by the different sensing layer using heuristic-based clustering approach. To
type of the nodes that sense different parameters. In such overcome this limitation, in the following section, the graph-
cases, formation of heterogeneous cluster using different based clustering algorithm is proposed. In this approach,
type of nodes and homogenous cluster using similar type clustering is performed by IoT layer instead of sensing layer,
of nodes helps to serve the particular functionality of an which is detailed in the next subsection.
application. Similar types of nodes or homogeneous nodes
are categorized based on the type of sensors or sensed values. 4.2. Graph-Based Clustering Approach. An energy-efficient
For example, consider the cluster of nodes in which all method for clustering is a need for IoT-based applications
nodes sense the temperature data and another cluster of to extend the network lifetime in real scenario. Assuming
nodes senses humidity value. Depending upon the need of that the location information of sensing nodes is available to
an application, the cluster may be activated and the data may the IoT nodes [46] and using this location information, IoT
be acquired for further processing. Even the same node may nodes can form the clusters and perform CH selection. As
have temperature and humidity sensors mounted on it. Such mentioned earlier, this strategy is also called top to bottom
nodes may be a part of both clusters with overlapping. approach, that is, from IoT layer to sensing layer. Generally,
In this context, for various applications in IoT [2, 4, IoT layer has high end processing capability as compared to
5, 39], two cases can be defined. In case 1, nodes want to ordinary nodes in the sensing layer. This strategy helps us
communicate with their own family type. For example, few deal with the issues of (i) optimal routing in sensing layer for
nodes are deployed to determine the light intensity and other communication across the clusters and (ii) less overhead on
few of them are for detection of carbon monoxide, which sensing devices for cluster formation and CH selection.
implies that the job descriptions of the nodes are different. In our proposed algorithm, the graph theory is used
In such cases, cluster formation and CH selection processes for clustering and for routing across the two layers, which
are applied to their own family type. CH is selected from provides an energy-efficient solution in the IoT environment
one among the same type of family. If a node broadcasts as evident from the simulation results. Graph theory has
packets within its range of transmission, the other family been used to solve many real-time problems in an optimal
Wireless Communications and Mobile Computing 9

2
Start
1 16
15
4 17
3 Input graph G(V, E)

14 18

13 20 Till all the nodes

5 For each
are considered in
vertex v ∈ V
19 the network
7 12
9
6
11
Select a node u with maximum degree
8 10 and residual energy
Figure 4: Example of dominating set.
Find the neighbors of node u

way [47]. In the proposed approach, dominating set and Add the node u to dominating set
bipartite graphs are utilized. Assume that graph 𝐺 = (𝑉, 𝐸) that acts as CH
is an undirected graph with 𝑉 as set of vertices and 𝐸 as
set of edges representing a network, where the vertices are
the sensor nodes and edges are the communication links in Update V by removing the covered nodes of u
the network, respectively. The proposed clustering algorithm and its neighbours from the existing vertices
initially divides 𝑉 into a collection of subsets {𝑉1 , 𝑉2 , . . . , 𝑉𝑘 }
but not inevitably disjoint, where V = ⋃𝑘𝑖=1 𝑉𝑖, such that each Dominating set acts as CHs
subset 𝑉𝑖 induces a connected subgraph of 𝐺. Overlapping communicate with the IoT nodes
of these induced subgraphs could be possible. Then, every
such subset vertex is a cluster. Here, k is the total number of
subgraphs in graph G. Stop
Let us say that G = (V, E) is an abstracted graph
constructed such that for every vertex V𝑖 ∈ V represents a Figure 5: Flowchart of graph-based algorithm.
subset 𝑉𝑖 . In general, in each cluster, a particular vertex is
selected, which acts as cluster head or cluster leader and that
2
corresponding subset forms the cluster. A graph 𝐺 = (𝑉, 𝐸)
1 16
is said to be a dominating set (DS), if it is a subset DS ⊆ 𝑉, 15
such that every vertex v ∈ V is either adjacent to a vertex of 4 17
DS or in DS. For instance, as shown in Figure 4, the sensing
3
layer has two types of color nodes indicating the CHs and
cluster members. The black color nodes are CHs, that is, 3, 14 18
7, 10, 14, and 18, which form a dominating set of the graph
because these vertices cover all the remaining white color 13 20
5
nodes. Further, in this approach, the residual energy of the
nodes is considered in the selection of dominating set. 7 12 19
The advantage of having residual energy as one of the 9
6
criteria for cluster head selection is that if these nodes are 11
selected as CHs in the network, the communication for
information collection from sensors and for controlling them 8 10
can be uniformly distributed to each and every vertex in just
one hop. The white nodes communicate with CHs and CHs Figure 6: Connected dominating set.
to IoT nodes and vice versa.
Dominating set problem is classical NP-complete deci-
sion problem [47] but many approximation algorithms exist, in Algorithm 2. The flowchart of the proposed graph-based
which provide optimal solutions up to a certain factor. A algorithm is shown in Figure 5.
simple approach is using the greedy strategy by selecting the If you connect all the vertices in the dominating set,
node that has highest degree with maximum residual energy then it is called connected dominating set, as shown in
and removes its neighbors from the set. The algorithm starts Figure 6. For routing computation in the network, connected
with empty DS and greedily adds nodes to DS until all nodes dominating sets are helpful. This routing application can be
are covered. The natural greedy approach for DS is given achieved by using a small connected dominating set that acts
10 Wireless Communications and Mobile Computing

Input Data: Graph 𝐺(𝑉, 𝐸)

Result: DS (Dominating set)
//Greedy algorithm for identifying the dominating set
(1) for each V ∈ 𝑉 do
(2) 𝑢 = Select(Δ(𝐺), max(RE)). //where Δ(𝐺) is maximum degree and RE is residual energy;
(3) Find 𝑁(𝑢); //Where 𝑁(𝑢) indicates is neighbors of node 𝑢 {𝑢 | 𝑢 ∈ 𝐺};
(4) DS = DS ∪ 𝑢. //where DS indicates dominating set and adding 𝑢 into set;
(5) 𝑉 = update(𝑉) {𝑉 | 𝑉/ ∈ 𝑁(𝑢), 𝑢}. //Remove from the covered nodes from the existing vertices;
(6) end
(7) CHs = DS. //Dominating set nodes acts as CHs.;

Algorithm 2: Graph-based clustering algorithm.

IoT nodes in layer 1

2 I1 I2 I3
1 16
15
4
17
3
14 18
13
5 20 ＃（1 ＃（2 ＃（3 ＃（4
12

19 Cluster head nodes in layer 2

9
6 7 Figure 8: Bipartite graph representation of two layers.
11
8 10

Figure 7: Cluster formation on layer 2 by IoT nodes. reconfiguring the positions of IoT nodes, even in dense
deployment of sensing devices. This helps in monitoring and
understanding the events occurring in the network.
as a backbone for communication. For example, nodes 3, 5, In the absence of layered architecture, the network is
7, 9, 10, 12, 18, and 20 act as the routing path of the network. a flat topology. In flat topology, to communicate from one
The remaining nodes that are not in the set of connected path node to the base station or from a node in one cluster to
can have the communication by passing messages through another node in another cluster, multihop communication is
neighbors. The cluster formation after applying dominating needed and this leads to a situation where the nodes nearer
set in the network is shown in Figure 7. to the base station die out faster than the remaining nodes
Because of the mobility of nodes, it is desirable to avoid [6, 34], whereas the proposed approaches take at most two
the same node to be elected as CH again and again. Besides, hop communications from node to CH and CH to IoT node,
the significant reason for using residual energy as another reducing the communication cost drastically to a greater
parameter is that the possible candidates for CHs are the extent. Further, the communication for CH selection and
members of the dominating set by default. It is easy to find a cluster formation is carried out for only two times. From the
node with maximum residual energy as a cluster head using discussion, it is clear that our proposed approach is simple
(2) to (5). and efficient.
Further, bipartite graph [48, 49] is used to represent the
communication between IoT nodes and CHs as shown in 5. Simulation Results
Figure 8. A bigraph or bipartite graph is a graph such that
the vertices are divided into two disjoint sets of classes, say 𝐼 For simulation, the Cooja simulator [32] is used. For eval-
and 𝐶. These two disjoint classes are independent sets, such uating algorithms in IoT environment, all communication
that every edge connects a vertex in 𝐼 to one in C, where should be performed using IoT stack, which is designed
𝐼 is a set of IoT nodes, that is, 𝐼 = {𝐼1 , 𝐼2 , . . . , 𝐼𝑛 }, and specifically for energy-constrained environment for Internet
𝐶 is set of CHs, that is, 𝐶 = {CH1 , CH2 , . . . , CH𝑛 }. There of Things [40–43]. IoT protocol stack is completely different
are many advantages of using bipartite graph representation from the normal WSN protocol stack [9, 10, 40]. The best part
for communication among the two layers. For instance, of the work is that the whole system is evaluated using Cooja
resource management issues can be handled optimally and simulator [32] for emulating actual IoT environment. All the
network communications can be balanced by periodically nodes are deployed randomly in the network. For simulation,
Wireless Communications and Mobile Computing 11

IoT node

Border
router
node

IoT node

Radio Heterogeneous
communications sensing nodes

Figure 9: Deployment of 15 sensing motes and 3 IoT motes in the network.

the radio transmission range of each node is set to 50 meters

and these nodes are mobile in nature. As shown in Figure 9, Virtual IPv6
the simulation is carried out using three types of motes, address of IoT
node 1
namely, sky-mote, wis-mote, and Z1 mote, respectively.
These motes are built-in feature of the Cooja simulator.
In Figure 9, 15 sensing nodes are deployed, which are in the
lower layer. Among these 15 nodes, five nodes are of the type Virtual IPv6
sky-mote, which are in green color, five nodes are of the type address of IoT
wis-mote, which are in purple color, and five nodes are of node 2
the type Z1-mote, which are in orange color. Similarly, in the
upper layer, three IoT nodes are deployed with IPv6 addresses
and can be seen as yellow-colored nodes. For the purpose
of displaying the sensing values on browser, border router
node is deployed, which is highlighted in blue color. This Virtual IPv6
border router node acts as a bridge between the IoT nodes address of IoT
node 3
and web page. The sensed information like temperature and
light intensity values from sensing layer is communicated to
IoT layer. The corresponding values of the readings are as
shown in Figure 10. By accessing the IPv6 address of IoT
nodes in the web browser, it is possible to view and analyze
the average sensed information across the scattered network.
In the literature, large numbers of clustering algorithms
Figure 10: Sensed values of the network by IoT nodes.
have been reported for wireless sensor network. For better
evaluation of the proposed algorithms, it is necessary to
compare the methods with relevant approaches reported
in the literature. So far, there is no method reported for
clustering in IoT environment in the literature. In such a case, A network, whether IoT or WSN, inherits common
the comparison seems to be difficult but there are standard characteristics in itself; that is, both are wireless in nature. But,
clustering algorithms available for WSN, which are very as pointed out earlier, clustering in the IoT-based network or
baseline approaches in nature and robust in performance. wireless sensor network has to address the same set of issues
Various variants of these baseline approaches have been for real-time applications. Hence, the evaluation criteria
modified for hierarchical network in the wireless sensor or performance parameters of clustering algorithm remain
network, even with mobility of the nodes. This leads us to the same for both networks. In the absence of IoT-based
compare our algorithms for better evaluation with existing clustering algorithms in literature, the proposed clustering
baseline standard algorithms, like LEACH and its variants algorithms are compared with standard and most widely used
with mobility, which is described later in this section. clustering methods in wireless sensor network using various
12 Wireless Communications and Mobile Computing

Comparative power consumption analysis Comparative cluster head analysis

3 35

2.5 30
2
Power (mW)

Number of cluster heads

25
1.5
20
1

0.5 15

0 10
5 10 15 20 25 30
Nodes 5

LEACH-mobile approach [20] Heuristic approach

0
CLUBS approach [21] Graph approach 15 30 50 100
Nodes
Figure 11: Comparative analysis of power consumption for cluster-
ing using different approaches. LEACH-mobile approach [20] Heurstic approach
CLUBS approach [21] Graph approach

Figure 12: Comparison for number of cluster heads with variation

parameters like number of hops, number of cluster heads, in density of nodes.
power consumption, and communication cost [19–31].
LEACH [11] is the standard algorithm and is widely
used for clustering in various applications of WSN. As LEACH-mobile [20] and CLUBS [21] approaches is illus-
discussed in the literature, there are other variants of LEACH trated as follows.
approach [19–22], which address the concept of mobility in The operation of LEACH is divided into rounds. Each
the case of WSN. In this context, our proposed approaches round begins with a setup phase when the clusters are
are compared and evaluated with LEACH-mobile approach organized, followed by a steady-state phase when data is
[20] and CLUBS approach [21], respectively. The simulation transferred from the nodes to the cluster head and onwards
is performed with varying number of nodes. Figure 11 shows to the base station. LEACH-mobile approach forms clus-
the power consumption at each node in the network using all ters by using a distributed algorithm, where nodes make
four algorithms. autonomous decisions without any centralized control. In
The dark blue color line indicates the result using LEACH-mobile approach, because of the random election of
LEACH-mobile approach [20] and light blue color line indi- cluster heads, the optimal number and distribution of cluster
cates the power consumption of CLUBS approach [21]. The heads cannot be ensured [50]. The above complete process
green color represents the output using heuristic-based clus- is performed in the static environment. But, as stated earlier,
tering approach and the yellow color shows the result using we also incorporate the mobility, which results into varying
graph-based clustering approach. The power consumption of number of nodes from each round to another, and hence the
30 nodes is calculated for our proposed approach and com- number of cluster heads is more in each round. This is mainly
pared with CLUBS [21] and LEACH-mobile [20] approaches due to mobility and randomness in electing the cluster heads.
as shown in Figure 11. From Figure 11, it is clear that the power In the proposed algorithms, the residual energy is used
consumption using our approaches, namely, heuristic- and for CH selection. The simulation results indicate that the
graph-based approaches, is very less as compared to CLUBS stability of the network is preserved, which implies that the
[21] and LEACH-mobile [20] approaches. It is observed that, whole network lifetime is extended. In Figure 12, it can be
for all 30 nodes, LEACH-mobile and CLUBS algorithms noticed that, for small number of nodes like 15 and 30,
consume more power than our proposed approaches. The the numbers of CHs are almost the same for the proposed
efficacy of the proposed approach it demonstrated. It is and existing LEACH-mobile and CLUBS approaches, but
important to notice that the graph-based approach consumes as the number of nodes increases to 50 and 100, proposed
less power than the other three approaches, because the algorithms have less number of CHs compared to LEACH-
clustering is carried out by the IoT layer which reduces the mobile and CLUBS approaches. Since LEACH-mobile and
overall overhead of sensing nodes. CLUBS approaches elect the CHs based on randomness and
Another criterion for evaluation is the number of CHs with certain probability, the number of CHs is more. LEACH-
elected in the network while clustering. It is preferred that mobile and CLUBS approaches do not consider the residual
the network should have maximum coverage with minimum energy in the process of CH election. But, in our proposed
CHs. Indirectly, it reduces overall communication cost and approaches, the residual energy is considered for CH election,
helps in prolonging the network lifetime. The comparative which is more practical approach and justified compared to
analysis of CHs selection using LEACH-mobile [20], CLUBS randomly choosing the CH for real-time application.
[21], and proposed algorithms is shown in Figure 12. The Also, in terms of network coverage, using our approaches,
reason behind the more number of CHs in the case of the minimum CHs are required as depicted in Figure 13.
Wireless Communications and Mobile Computing 13

CHs: 1, 5, 6, 7, 9. CHs: 3, 10, 13, 28, 33, 65, 68, 82, 87.

(a) Number of CHs with 15 nodes (b) Number of CHs with 100 nodes

Figure 13

Worst case comparative analysis of number of hops Figure 14 depicts the result related to the number of hops
16
occurring while communicating to destination and base
14 station in the network using our proposed methods and
LEACH-mobile [20] and CLUBS [21] approaches.
12 From the figure, it is clear that the number of hops
required using our proposed methods is less compared to
Number of hops

10
the other two approaches, LEACH-mobile and CLUBS. The
8 reason is that LEACH-mobile approach operates with fixed
6
and single base station and the number of hops to reach the
base station from all the cluster heads is more. Considering
4 the worst case, the cluster heads are randomly selected,
which are farther from the base station; then, this may even
2
result in an increased number of hops to reach the base
0 station. However, in our proposed approach, due to two-layer
15 30 50 100 architecture, all nodes are able to reach the upper layer, that
Nodes is, IoT nodes that act as base station, in one hop only. The
LEACH-mobile approach [20] Heuristic approach
comparison is demonstrated for different set of nodes in the
CLUBS approach [21] Graph approach network. From Figure 14, the worst case scenario of choosing
number of hops is carried out and it is clear that the numbers
Figure 14: Worst case comparison for number of hops with variation of hops are 2, 4, 9, and 16 for LEACH-mobile approach and
in density of nodes. 3, 4, 8, and 15 for CLUBS approach for the 15, 30, 50, and
100 numbers of nodes, respectively. It is evident from the
above discussion that our proposed approaches outperform
the existing approaches.
For instance, as shown in Figure 13(a), in clustering and Another important parameter for evaluating the pro-
CH election with 15 nodes, nodes 1, 5, 6, 7, and 9 are elected as posed algorithms compared to other methods reported in the
CHs, whereas, in the same network, assuming that 100 nodes literature is overall communication cost. The communication
are deployed, 9 CHs are elected covering all the nodes, that cost using two-layer topology and flat topology is compared.
is, 3, 10, 13, 28, 33, 65, 68, 82, and 87, which is depicted in In typical WSN, the cost of communication overhead is
Figure 13(b). Further, the proposed algorithm is evaluated 50–70% more than the cost of computation [44, 45]. Lesser
in terms of the number of hops. As shown in Figure 14, number of communications leads to increasing the network
irrespective of number of nodes, the number of hops is one. lifetime. Since our proposed approach has less number of
This is because cluster formation is carried out using the CHs and less number of hops, this leads to reduced number
node transmission range. All the cluster members are able to of communications.
communicate to CH in a single hop. As shown in Figure 15, the proposed approaches ensure
Another comparison among these four algorithms per- less communication cost, which leads to prolonged network
formed is based on the number of hops required to reach lifetime. The stability in the network depends on the battery
the destination, that is, the base station in the given network. of the sensors. More communication implies that sooner the
14 Wireless Communications and Mobile Computing

Comparative communication cost analysis efficient communication to have data availability anywhere
350
and anytime with prolonged lifetime of the network. Two
300 clustering algorithms have been proposed using the neigh-
bors count and residual energy. The heuristic- and graph-
Number of communications

250 based approaches allow performing the clustering in bottom

to top and top to bottom ways depending on the need of an
200 IoT application. The proposed algorithms are evaluated using
a number of standard parameters and compared with existing
150 algorithms. The proposed layered IoT framework provides
hierarchical structure for efficient connectivity. Our proposed
100
clustering algorithms also allow forming different clusters
50 in presence of mobility and heterogeneity. The dynamic
cluster formation and efficient cluster head selection ensure
0 effective connectivity and efficient communication. From the
15 30 50 100
simulation results, it is clear that the proposed clustering
Nodes algorithms outperform the existing algorithm in terms of
LEACH-mobile approach [20] power consumption and the number of hops required for
CLUBS approach [21] communication across the network. The number of neighbor
Heuristic approach nodes and residual energy are proven to be very effective
Graph approach for cluster formation and cluster head selection. The graph-
based approach with dominant set and bipartite graph allows
Figure 15: Comparison of communication cost with variation in
density of nodes. forming a cluster optimally and also provides a routing path
for communication.
The algorithm implementations are carried out using
IoT-based Cooja simulator, which ensures that the proposed
network becomes unstable due to battery drain of the node. algorithms are evaluated in IoT environment. Comparative
It is observed that the number of communications in the case analysis of communication cost with respect to the flat
of LEACH-mobile [20] and CLUBS [21] approaches is sig- topology and layered topology draws to the conclusion that
nificantly more than our proposed approaches. Since in our layered topology provides better results. The crucial factor of
proposed approach we use only two times communication the network depends on the residual energy of the nodes,
for CH election and cluster formation, the communication which is taken into consideration for cluster formation,
cost is very less, whereas, in the LEACH-mobile and CLUBS which concludes that the proposed approaches are feasible.
approaches, they form the cluster using broadcasting of Results indicate that the optimal number of CHs has been
random messages for minimum three times, which results in elected and thus it indicates maximum network coverage.
more communication. From the above discussion, it can be Minimum number of communications and the amount
concluded that the proposed approaches are much better in of power consumption in the network promise prolonged
terms of communication cost and result in more stability or network lifetime. The proposed algorithm can be extended
prolonged lifetime of the network. to connect with IoT cloud to realize the computation in
fog environment. Various parameters can be incorporated
6. Conclusion to define the characteristics of heuristic or graph edges for
clustering depending on the applications on hand.
Due to advancement in the technology and need for machine-
to-machine connectivity, the Internet of Things overplays the Conflicts of Interest
role compared to wireless sensor network. In this context,
different applications based on IoT need to be executed The authors declare that there are no conflicts of interest
efficiently in terms of energy and communication. To achieve regarding the publication of this article.
this, there is a need to collaborate among various devices at
various levels, which can be achieved by the grouping of these Acknowledgments
devices, that is, through the clustering. In this view, we intro-
duced a pragmatic architectural framework for IoT in this This work is supported by the Ministry of Electronics and
paper and subsequently proposed two clustering algorithms, Information Technology (MeitY), funded by Ministry of
that is, heuristic- and graph-based algorithms, respectively. Human Resource Development (MHRD), Government of
The proposed two-layer framework is justified as most of India (Grant no. 13(4)2016-CC&BT).
the applications using IoT are sensing based static or dynamic
using IP-based or non-IP-based devices. Secondly, the two-
layer architecture allows one to have the sensed values or
References
information anywhere and anytime using IoT devices. For [1] J. S. Kumar and D. R. Patel, “A survey on internet of things:
any IoT application having huge number of nodes at sensing security and privacy issues,” International Journal of Computer
layer and IoT layer, it must have effective connectivity and Applications, vol. 90, no. 11, pp. 20–26, 2014.
Wireless Communications and Mobile Computing 15

[2] M. A. Zaveri, S. K. Pandey, and J. Sathish Kumar, “Collaborative [18] H. Le, D. Hoang, and R. Poliah, “S-Web: An efficient and self-
service oriented smart grid using the Internet of Things,” in Pro- organizing wireless sensor network model,” in Springer Lecture
ceedings of IEEE at International conference on communication Notes in Computer Science, vol. 5186, p. 179, 2008.
and signal processing, pp. 1412–1418, April 2016. [19] G. S. Kumar, V. P. Mv, and K. P. Jacob, “Mobility metric based
[3] S. Penga, W. Yao-huib, and C. Chaoc, “IoT core technology and LEACH-Mobile protocol,” in Proceedings of 16th International
its applications,” Journal of Communications Technology, vol. 5, Conference on Advanced Computing and Communications, pp.
p. 35, 2011. 248–253, India, 2008.
[4] K. S. Yeo, M. C. Chian, T. C. W. Ng, and D. A. Tuan, “Internet [20] G. Renugadevi and G. M. Sumithra, “An analysis on LEACH-
of things: Trends, challenges and applications,” in Proceedings mobile protocol for mobile wireless sensor networks,” Interna-
of the 14th International Symposium on Integrated Circuits, ISIC tional Journal of Computer Applications, vol. 65, no. 21, 2013.
2014, pp. 568–571, December 2014. [21] R. Nagpal and D. Coore, “An algorithm for group formation
[5] J. Gubbi, R. Buyya, S. Marusic, and M. Palaniswami, “Internet in an amorphous computer,” in Proceedings of the 10th Inter-
of Things (IoT): a vision, architectural elements, and future national Conference onparallel and distributed systems, LV, NV,
directions,” Future Generation Computer Systems, vol. 29, no. 7, USA, October 1998.
pp. 1645–1660, 2013.
[22] T. Firdaus and M. Hasan, “A survey on clustering algorithms for
[6] A. A. Abbasi and M. Younis, “A survey on clustering algorithms energy efficiency in wireless sensor network,” in Proceedings of
for wireless sensor networks,” Journal of Computer communica- the 3rd International Conference on Computing for Sustainable
tionsin Elsevier, vol. 30, pp. 2826–2841, 2007. Global Development, INDIACom 2016, pp. 759–763, March 2016.
[7] R. Khan, S. U. Khan, and R. Zaheer, “Future internet: the [23] F. D. Tolba, W. Ajib, and A. Obaid, “Distributed clustering
internet of things architecture, possible applications and key algorithm for mobile wireless sensors networks,” in Proceedings
challenges,” in Proceedings of the FIT, pp. 257–260, December of the IEEESENSORS, pp. 1–4, 2013.
2012.
[24] D.-S. Kim and Y.-J. Chung, “Self-organization routing protocol
[8] “Architecturalreference model (ARM) for internet of things,” supporting mobile nodes for wireless sensor network,” in Pro-
http://iotforum.org/wp-content/uploads/2014/09/120613-IoT- ceedings of the 1st International Multi-Symposiums on Computer
A-ARM-Book-Introduction-v7.pdf, 2011. and Computational Sciences (IMSCCS ’06), vol. 2, pp. 622–626,
[9] J. S. Kumar and M. A. Zaveri, “Clustering for collaborative June 2006.
processing in IoT network,” in Proceedings of the SecondInterna-
[25] P. Ding, J. A. Holliday, and A. Celik, “Distributed energy-
tional Conference on IoT in Urban Space, ACM, pp. 95–97, 2016.
efficient hierarchical clustering for wireless sensor networks,” in
[10] J. S. Kumar and M. A. Zaveri, “Hierarchical Clustering for Proceedings of the IEEE International Conference on Distributed
Dynamic and Heterogeneous Internet of Things,” Elsevier Pro- Computing in Sensor Systems (DCOSS’05), 2005.
cedia Computer Science, vol. 93, pp. 276–282, 2016.
[26] J. Lloret, M. Garcia, D. Bri, and J. R. Diaz, “A cluster-based
[11] H. Gou and Y. Yoo, “An energy balancing LEACH algorithm for architecture to structure the topology of parallel wireless sensor
wireless sensor networks,” in Proceedings of the 7th International networks,” Sensors, vol. 9, no. 12, pp. 10513–10544, 2009.
Conference on Information Technology: New Generations (ITNG
’10), pp. 822–827, April 2010. [27] S. Rani, R. Talwar, J. Malhotra, S. H. Ahmed, M. Sarkar, and
H. Song, “ICCBP: Inter cluster chain based protocol with
[12] V. Loscri, G. Morabito, and S. Marano, “A two-levels hierarchy cross layer interaction for randomly deployed wireless sensor
for low-energy adaptive clustering hierarchy (TL-LEACH),” in networks,” Ad Hoc & Sensor Wireless Networks, vol. 36, no. 1-4,
Proceedings of the IEEE Vehicular Technology Conference, vol. 3, pp. 257–284, 2017.
p. 1809, 2005.
[28] A. Manjeshwar and D. P. Agrawal, “TEEN: a routing protocol for
[13] S. Bandyopadhyay and E. J. Coyle, “An energy efficient hier-
enhanced efficiency in wireless sensor networks,” in Proceedings
archical clustering algorithm for wireless sensor networks,” in
of the 15th International Parallel and Distributed Processing
Proceedings of Twenty-Second Annual Joint Conference of the
Symposium (IPDPS ’01), 2001.
IEEE Computer and Communications. IEEE Societies, vol. 3, pp.
1713–1723, 2003. [29] A. Manjeshwar and D. P. Agrawal, “APTEEN: a hybrid protocol
for efficient routing and comprehensive information retrieval
[14] C. Li, M. Ye, G. Chen, and J. Wu, “An energy-efficient unequal
in wireless sensor networks,” in Proceedings of the IEEE at
clustering mechanism for wireless sensor networks,” in Pro-
the International Parallel and Distributed Processing Symposium
ceedings of IEEE International Conference on Mobile Adhoc and
(IPDPSı́02), pp. 1530–2075, 2002.
Sensor Systems Conference, pp. 1–8, 2005.
[15] S. Yi, J. Heo, Y. Cho, and J. Hong, “PEACH: power-efficient [30] J. Kim, S. Park, Y. Han, and T. Chung, “CHEF: Cluster Head
and adaptive clustering hierarchy protocol for wireless sensor Election mechanism using Fuzzy logic in wireless sensor net-
networks,” Computer Communications, vol. 30, no. 14-15, pp. works,” in Proceedings of the ICACT, pp. 654–659, 2008.
2842–2852, 2007. [31] H. Bagci and A. Yazici, “An Energy Aware Fuzzy Unequal Clus-
[16] B. Gong, L. Li, S. Wang, and X. Zhou, “Multihop routing tering Algorithm for wireless Sensor Networks, Fuzzy Systems
protocol with unequal clustering for wireless sensor networks,” (FUZZ),” in Proceedings of the IEEE International Conference on
in Proceedings of IEEE and ISECS International Colloquium Digital Object Identifier, pp. 1–8, 2010.
on Computing, Communication, Control, and Management, [32] “Cooja simulator on contiki2.7 operating system,” http://www
CCCM’08, vol. 2, pp. 552–556, August 2008. .Contiki-os.org, 2013.
[17] O. Younis and S. Fahmy, “HEED: a hybrid, energy-efficient, [33] P. Kuila and P. K. Jana, “Energy efficient clustering and rout-
distributed clustering approach for ad hoc sensor networks,” ing algorithms for wireless sensor networks: particle swarm
IEEE Transactions on Mobile Computing, vol. 3, no. 4, pp. 366– optimization approach,” Engineering Applications of Artificial
379, 2004. Intelligence, vol. 33, pp. 127–140, 2014.
16 Wireless Communications and Mobile Computing

[34] O. Boyinbode, H. Le, and M. Takizawa, “A survey on clustering

algorithms for wireless sensor networks,” International Journal
of Space-Based and Situated Computing, vol. 1, no. 2-3, pp. 130–
136, 2011.
[35] D. S. Mantri, N. R. Prasad, and R. Prasad, “Mobility and Het-
erogeneity Aware Cluster-Based Data Aggregation for Wireless
Sensor Network,” Journal of Springer, vol. 86, no. 2, pp. 1–19,
2015.
[36] A. M. Krishnan and P. G. Kumar, “An effective clustering
approach with data aggregation using multiple mobile sinksfor
heterogeneous wsn, Wireless Personal Communications,” Jour-
nal of Springer, pp. 1–12, 2015.
[37] H. Karvonen, J. Suhonen, J. Petäjäjärvi, M. Hämäläinen, M.
Hännikäinen, and A. Pouttu, “Hierarchical architecture for
multi-technology wireless sensor networks for critical infras-
tructure protection,” Wireless Personal Communications, vol. 76,
no. 2, pp. 209–229, 2014.
[38] S. Jabbar, A. A. Minhas, A. Paul, and S. Rho, “Multilayer cluster
designing algorithm for lifetime improvement of wireless sen-
sornetworks,” The Journal of Supercomputing, vol. 70, no. 1, pp.
104–132, 2014.
[39] L. Atzori, A. Iera, and G. Morabito, “The internet of things: a
survey,” Journal of Computer networks in Elsevier, vol. 54, no. 15,
pp. 2787–2805, 2010.
[40] M. R. Palattella, N. Accettura, X. Vilajosana et al., “Standardized
protocol stack for the internet of (important) things,” IEEE
Communications Surveys & Tutorials, vol. 15, no. 3, pp. 1389–
1406, 2013.
[41] N. Salman, I. Rasool, and A. H. Kemp, “Overview of the IEEE
802.15.4 standards family for low rate wireless personal area
networks,” in Proceedings of 7th International Symposium on
Wireless Communication Systems (ISWCS), pp. 701–705, UK,
2010.
[42] Z. Shelby and C. Bormann, 6LoWPAN: The wireless embedded
Internet, vol. 43, John Wiley and Sons, 2011.
[43] Z. Shelby, K. Hartke, and C. Bormann, “The constrained ap-
plication protocol (coap),” https://tools.ietf.org/html/rfc7252,
2014.
[44] W. Si, H.-Y. Shi, and P. Liu, A Collaborative Data Gathering
Mechanism Based on Fuzzy Decision for Wirelesssensor Net-
works, vol. 11 of SAGE Publications Sage, 2015.
[45] S. Funke, A. Kesselman, U. Meyer, and M. Segal, “A simple
improved distributed algorithm for minimum CDS in unit disk
graphs,” ACM Transactions on Sensor Networks, vol. 2, no. 3, pp.
444–453, 2006.
[46] S. K. Pandey and M. A. Zaveri, “Quasi random deployment and
localization in layered framework for the Internet of Things,” in
The Computer Journal, vol. 61, pp. 1–21, Oxford university press,
2017.
[47] M. R. Garey and D. S. Johnson, Computers and Intractability: A
Guide to The Theory of NP-Completeness, WH Free. Co., San Fr,
San Fr, Calif, USA, 1979.
[48] A. S. Asratian, T. M. Denley, and R. Haggkvist, Bipartite Graphs
and Their Applications, vol. 131, Cambridge University Press,
1998.
[49] J. E. Hopcroft and R. M. Karp, “An n5 /2 Algorithm for Max-
imum Matchings in Bipartite Graphs,” SIAM Journal on Com-
puting, vol. 2, no. 4, pp. 225–231, 1973.
[50] W. R. Heinzelman, J. Kulik, and H. Balakrishnan, “Adaptive
protocols for information dissemination in wireless sensor
networks,” in Proceedings of the ACM MobiCom, 1999.
 International Journal of

Rotating Advances in
Machinery Multimedia

The Scientific
Engineering
Journal of
Journal of

Hindawi
World Journal
Hindawi Publishing Corporation Hindawi
Sensors
Hindawi Hindawi
www.hindawi.com Volume 2018 http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of

Control Science
and Engineering

Advances in
Civil Engineering
Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at

www.hindawi.com

Journal of
Journal of Electrical and Computer
Robotics
Hindawi
Engineering
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

VLSI Design
Advances in
OptoElectronics
International Journal of

International Journal of
Modelling &
Simulation
Aerospace
Hindawi Volume 2018
Navigation and
Observation
Hindawi
www.hindawi.com Volume 2018
in Engineering
Hindawi
www.hindawi.com Volume 2018
Engineering
Hindawi
www.hindawi.com Volume 2018
Hindawi
www.hindawi.com www.hindawi.com Volume 2018

International Journal of
International Journal of Antennas and Active and Passive Advances in
Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration
Hindawi Hindawi Hindawi Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018