Vol.:(0123456789) 1 3
Journal of Ambient Intelligence and Humanized Computing
https://doi.org/10.1007/s12652-020-02599-3
ORIGINAL RESEARCH
Predicting the energy consumption in software defned wireless
sensor networks: a probabilistic Markov model approach
Atefeh Rahimifar
1
· Yousef Seif Kavian
1
· Hooman Kaabi
1
· Mohammad Soroosh
1
Received: 18 January 2019 / Accepted: 3 October 2020
© Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract
The smart world is connecting all universe more than ever thought possible, benefting from the signifcant advances of the
Internet of Things (IoT) applications using wireless sensor networks (WSN) as the core technology. A challenging issue in the
IoT paradigm is the heterogeneity in diferent parts of the network. The network developers need to use resources belonging
to diferent platforms for their applications, and the software defned network (SDN) approach is a mainly considered solu-
tion. In this paper, a software defned wireless sensor network (SDWSN) with an energy predictor model (SDWSN-EPM)
based on the Markov probabilistic model is proposed to reduce the energy consumption and the network latency. The energy
consumption rate (ECR) of the sensor nodes is modeled using the Markov model and the states of the sensor nodes. The ECR
is used by the SDN controller to predict the residual energy level of the nodes and consequently, the energy consumption
of the network. The cumulative distribution functions (CDF) of the delay, power consumption, and the network lifetime in
both SDWSN and SDWSN-EPM schemes are compared. The results confrm that the SDWSN-EPM model signifcantly
improves the performance of the sensor networks.
Keywords Software defned wireless sensor networks · Energy consumption prediction · Markov model · Performance
evaluation · Internet of things
1 Introduction
The Internet of Things (IoT) is rapidly taking over the world.
Smart healthcare, smart grid, smart cities, smart homes,
smart vehicles, smart underwater networks, and intelligent
transportation are some examples of the IoT applications
which are making signifcant development in our daily life
(Ma and Li 2020; Yamauchi et al. 2020; Menon et al. 2020).
Wireless sensor network (WSN) is an inseparable part of the
IoT that collect and transfer information. Moreover, with
advancements in IoT, wireless devices can be integrated into
existing wired systems, making it a heterogeneous network,
where diferent mediums, protocols, and software parts may
coexist in the same network (Cecílio et al. 2017). Therefore,
creating a seamless interaction between heterogeneous net-
works is a signifcant challenges of the IoT.
In recent years a new achievement has been proposed
to solve IoT problems such as management difculty, het-
erogeneity, and low fexibility of networks (Anadiotis et al.
2016, 2018). This approach, which applies wireless sensor
nodes based on the SDN paradigm and OpenFlow (McK-
eown et al. 2008), is called software defned wireless sensor
network (SDWSN). The SDWSN separates the data plane
from the control plane. In the data plane, wireless sensor
nodes transmit and handle incoming packets based on their
fow tables (McKeown et al. 2008). The control plane, on the
other hand, contains one or more controllers that decide and
determine the policies of the sensor network. That communi-
cates with the data plane via the southbound interface. Fur-
thermore, the control plane can be connected to the applica-
tion layer via the northbound application program interface
(API) (R. Wang et al. 2018; Modieginyane et al. 2018). In
this paper, the main motivation of using the SDWSN is to
create a fexible and scalable wireless network with easy
management of IoT infrastructure.
Nevertheless, the SDWSNs are immature as there are
many challenges such as optimization, standardization of
control and data plane, security, and energy consumption.
* Yousef Seif Kavian
y.s.kavian@scu.ac.ir
1
Faculty of Engineering, Shahid Chamran University
of Ahvaz, Ahvaz, Iran