IEEE Network - March / April 2017 - page 92

IEEE Network • March/April 2017
90
0890-8044/17/$25.00 © 2017 IEEE
A
bstract
Fifth generation (5G) wireless networks adopt
the deployment of ultra-dense small cells for effi-
cient slicing of radio resources. This conceptu-
al change in network structure aims to meet the
rapid increase in mobile data traffic and connect-
ed devices. However, limited free spectrum and
dynamic assignment of resources are main con-
cerns when considering the cognitive small cells
solution. Therefore, there is a need to map traffic
patterns with the number of cognitive small cells
to provide an optimized network architecture
operating with adequate spectrum resources. This
article investigates the case when network densifi-
cation exceeds the radio resource capacity, caus-
ing a large scale overlapping in cell coverage area
and used channels. Taking into consideration cog-
nitive network performance characteristics, we
identify two spectrum coexistence frameworks,
Space Filling
and
Time Filling
, to improve spectrum
utilization and scalability for moderately large net-
works. Simulations show that there is a turning
point when network performance starts to decline
as the number of cognitive small cells exceeds
the shared resources in a site area, subject to a
certain load profile. This optimization of network
structure, based on spectrum transmission oppor-
tunities, brings about a new topic for operators
and research communities considering small cells
operating in the unlicensed band.
I
ntroduction
Cognitive wireless networks enable dynamic
access to underutilized spectrum in the licensed
band. This type of network provides higher values
for data rates and capacity, as users tend to down-
load data instead of making voice calls, using
enhanced radio access and decision making tech-
niques [1]. Spectrum efficiency is determined by
a variety of parameters such as successful assess-
ment of channel availability, selection of suitable
and short transmission links, and traffic distribution
between site access points. Therefore, user associ-
ation and access point selection are key features
for ultra-dense radio networks employing distrib-
uted small cells. Specifically, these features can
reduce the impacts of overlapped coverage areas,
and also determine the instantaneous power of
access points in multi-shaped coverage areas as
provided by the IEEE 802.22 standard for cogni-
tive wireless regional area networks (WRANs) [2].
The assumptions of spectrum shortage and the
high-cost of deploying access points, in known
network design paradigms, have impacted the
network structure of the current cellular systems
such as fourth generation (4G) Long Term Evolu-
tion (LTE) standards [3]. Therefore, LTE networks
are designed with relatively large sized coverage
areas to operate at the upper bound limits of
spectrum efficiency [4]. This leaves less margin for
any new developments using the licensed band
while motivating the dense deployment of small
cells operating in the unlicensed band.
In cognitive radio communications, the small
cell solution is driven by the theoretical approach
of intensive network slicing provided with new
spectrum coexistence techniques for efficient spec-
trum utilization [5], as shown in Fig. 1. This also
requires efficient spectrum sensing mechanisms
and instant sharing of sensor information between
contiguous small cells [6]. Moreover, small cells
can improve users’ connectivity in heterogeneous
networks (HetNets) through shorter wireless links
of optimal pilot power. Currently, wireless net-
works employ predefined converged small cells
[7], without considering the challenges that may
emerge when deploying a large number of sim-
ilar cells at the same sites. Cognitive radio small
cells impose additional challenges because of their
opportunistic spectrum access and flexible cov-
erage areas. Therefore, mobile operators need to
consider new spectrum coexistence models when
sharing the spectrum with other users regardless of
being the owners of the spectrum band.
Resource allocation provides users with the
necessary spectrum considering the requested
quality of service (QoS), assigns users between
different tiers, and enables high volumes of data
delivery to end users. Most of the resource allo-
cation approaches for cooperative cognitive radio
networks are presented without much consider-
ation for overlapped small cell domains or shared
use of channels between cognitive radio users
[8]. In fact, the trend in the literature seems to
be focused on improving the allocation of radio
resources in cellular networks consisting of over-
laid small cells to deal with the challenge of lim-
ited backbone capacity. In the cognitive radio
context, this may happen when small cells release
channels back to the main network or neighbored
cells as soon as they fulfill a transmission request.
To achieve this, cognitive small cells needs to
employ an efficient medium access control (MAC)
mechanism that can adaptively access available
channels without interfering with the surround-
ing wireless environment. Both network planning
and spectrum availability need to be combined
to efficiently utilize the scarce radio resources by
deploying the optimal number of cognitive small
cells. This defines why we need to jointly consid-
er the impact of small cell deployment and the
resource allocation problem in two-tire networks,
rather than separately.
Planning of Ultra-Dense Wireless Networks
Anwer Al-Dulaimi, Saba Al-Rubaye, John Cosmas, and Alagan Anpalagan
ACCEPTED FROM OPEN CALL
Digital Object Identifier:
10.1109/MNET.2017.1500258NM
1...,82,83,84,85,86,87,88,89,90,91 93,94,95,96,97,98,99,100
Powered by FlippingBook