sensors

Communication
Dynamic Spectrum Sharing for Future LTE-NR Networks †
Gordana Barb, Florin Alexa * and Marius Otesteanu

Department of Communications, Politehnica University Timis, oara, 300006 Timis, oara, Romania;
gordana.barb@student.upt.ro (G.B.); marius.otesteanu@upt.ro (M.O.)
* Correspondence: florin.alexa@upt.ro
† This manuscript is extension version of the conference paper: Barb, Gordana, Marius Otesteanu, and Marius
Roman. “Dynamic Spectrum Sharing for LTE-NR Downlink MIMO Systems.” In Proceedings of the 2020
International Symposium on Electronics and Telecommunications (ISETC), Timisoara, Romania,
5–6 November 2020; pp. 1–4.

Abstract: 5G is the next mobile generation, already being deployed in some countries. It is expected
to revolutionize our society, having extremely high target requirements. The use of spectrum is,
therefore, tremendously important, as it is a limited and expensive resource. A solution for the
spectrum efficiency consists of the use of dynamic spectrum sharing, where an operator can share
the spectrum between two different technologies. In this paper, we studied the concept of dynamic
spectrum sharing between LTE and 5G New Radio. We presented a solution that allows operators to
offer both LTE and New Radio services using the same frequency bands, although in an interleaved
mode. We evaluated the performance, in terms of throughput, of a communication system using the
dynamic spectrum sharing feature. The results obtained led to the conclusion that using the dynamic
spectrum sharing comes with a compromise of a maximum 25% loss on throughput. Nevertheless,
the decrease is not that substantial, as the mobile network operator does not need to buy an additional



15 MHz of bandwidth, using the already existing bandwidth of LTE to offer 5G services, leading to

 cost reduction and an increase in spectrum efficiency.
Citation: Barb, G.; Alexa, F.;
Otesteanu, M. Dynamic Spectrum Keywords: dynamic spectrum sharing; LTE; 5G NR; throughput; spectrum efficiency
Sharing for Future LTE-NR Networks.
Sensors 2021, 21, 4215. https://
doi.org/10.3390/s21124215
1. Introduction
Academic Editor: Dimitrie
The next mobile generation following Long Term Evolution (LTE) that has started its
C. Popescu
deployment in some countries is the 5th Generation (5G). The main focus of LTE was set
to increase data transfer rates, while 5G is expected to revolutionize our society, focusing
Received: 10 May 2021
Accepted: 16 June 2021
not only on delivering extreme mobile broadband, but also in the fields of critical machine
Published: 19 June 2021
communication and massive machine communication. New applications will emerge,
and the target values and requirements proposed are extremely demanding [1]. The main
Publisher’s Note: MDPI stays neutral
challenges of 5G systems consist of increasing data transfer rates, reducing latency, and
with regard to jurisdictional claims in
increasing capacity, spectrum efficiency, and network energy efficiency, which will be
published maps and institutional affil- necessary for different application scenarios [2].
iations. The current network architecture cannot sustain all the requirements and target val-
ues of 5G New Radio (NR). Therefore, the 3rd Generation Partnership Project (3GPP)
released two variations for the new network architecture of 5G NR communication sys-
tems: Non-standalone (NSA) and Standalone (SA) [3]. The main difference between both
Copyright: © 2021 by the authors.
types is that the NSA architecture is based and depends on the LTE core network, while
Licensee MDPI, Basel, Switzerland.
the SA architecture uses a novel next-generation core network, not depending of any
This article is an open access article
LTE infrastructure.
distributed under the terms and The Internet of Things (IoT), mobile internet, and Cognitive Radio (CR) stand as
conditions of the Creative Commons relevant driving forces for 5G development [4–7]. The IoT technology has the potential
Attribution (CC BY) license (https:// ability to connect almost everything to the internet, which will lead to the massive growth
creativecommons.org/licenses/by/ of devices that require network acceptance. Particularly, in 2018, there were approximately
4.0/). 22 billion connected devices, which is the equivalent of around 2.9 devices/person. It

Sensors 2021, 21, 4215. https://doi.org/10.3390/s21124215 https://www.mdpi.com/journal/sensors

Sensors 2021, 21, 4215 2 of 15

is expected that by 2025, the number of connected devices to be nearly 38.6 billion [8].
Simultaneously, the mobile internet has surpassed the traditional mobile communications,
offering users innovative experiences. On the other hand, cognitive radio represents
a technology that has evolved in order to improve the spectrum usage, exploiting both
licensed and unlicensed bands. It became a promising wireless communications technology
that is able to solve spectrum problems by: observation, learning, and intelligent adaptation
to obtain an optimal frequency band [9].
In addition to the network architecture, another major challenge of 5G systems consists
of the available frequency spectrum. Presently, the spectrum used for mobile communi-
cations is located between 300 MHz and up to 6 GHz. Due to the fact that most of it is
completely saturated, new frequency bands above 6 GHz need to be taken into consider-
ation. Hence, it is expected that 5G NR systems will use frequency bands above 6 GHz,
reaching up to 300 GHz [10]. Therefore, the main frequency spectrum to be used for 5G
systems can be in licensed, shared, and unlicensed bands, see Table 1. The low-band spec-
trum has the goal of supporting high coverage and penetration; the mid-band spectrum
provides data transfer rates up to 1 Gbps; the high-band spectrum supports peak data
transfer rates up to 20 Gbps. The last group, even though it supports such high data rates,
comes with the inconvenience of low coverage and penetration, being indicated to be
applied in closed/indoor environments.

Table 1. 5G spectrum options.

Frequency Band Characteristics Spectrum Use

Wide area coverage
600–2600 MHz Spectrum sharing
Reframing from
old technologies
Mid-band mainstream for 5G
3.3–5 GHz
Massive MIMO-crowded and
urban areas Independent deployment
Unlicensed spectrum (without old technologies)
5–6 GHz
No spectrum
licensed necessary
Allocated mmWaves for 5G
24–39 GHz
Increased data rates

It is clear that the frequency spectrum is a scarce and limited resource that constitutes
an important factor in mobile communication systems, as well as the related cost for the
Mobile Network Operator (MNO). In this context, new spectrum explorations [11,12],
higher energy efficiency [13], and dynamic spectrum usage [14–18] have become the new
features of communication networks.
The topic of spectrum sharing in the bands of old communication systems started
drawing the attention of researchers, as it is the safest and most economical solution [19].
The standardization procedure for the spectrum sharing principles started in March 2017
by 3GPP. One of the solutions presented regarding the spectrum allocation for 5G NR
systems comprises the use of the existing frequency spectrum used by the already deployed
mobile generations.
Spectrum sharing is based on the flexibility of the physical layer and the fact that in
the LTE network, all channels are assigned in the time-frequency domain. This way, the
flexibility of the 5G NR radio interface can be used for reference signals, allowing dynamic
configuration and minimizing collisions between NR and LTE during simultaneous data
transmission. Consequently, there is the possibility of sharing a frequency domain within the
same communication channel. A comprehensive overview on the different ways of spectrum
sharing that has been investigated in recent years is found in [14]. In addition, in [20,21],
new schemes and algorithms for dynamic spectrum sharing between Global System for
Mobile Communications (GSM) and LTE technologies were investigated. Regarding the
IoT, spectrum sharing is a preferable approach to cope with the conflicts between massive
 

between massive IoT connections and limited spectrum resources as it was discussed in
Sensors 2021, 21, 4215 [22–25]. It can also be used to solve vertical requirements and the competition in the ac‐
3 of 15
quisition of frequency bands between MNOs [26,27], as well as to improve spectrum uti‐
lization in Cognitive Radio (CR) and TV white space [28–30]. The leading mobile produc‐
ers have shown massive interest in developing solutions for dynamic spectrum sharing.
IoT connections and limited spectrum resources as it was discussed in [22–25]. It can also
These are presented in Table 2. 
be used to solve vertical requirements and the competition in the acquisition of frequency
bands between MNOs [26,27], as well as to improve spectrum utilization in Cognitive
Table 2. Leading producers and adopted DSS solution. 
Radio (CR) and TV white space [28–30]. The leading mobile producers have shown massive
interest in developing solutions for dynamic spectrum sharing. These are presented in
Leading Producers  Name of DSS Solution 
Table 2.
Ericsson  ESS—Ericsson Spectrum Sharing 
Huawei 
Table 2. Leading producers and adopted DSS solution. CloudAIR 
MediaTeck 
Leading Producers
Dimensity 1000 
Name of DSS Solution
Nokia  Dynamic Spectrum Sharing (DSS) 
Ericsson ESS—Ericsson Spectrum Sharing
QualComm 
Huawei Snapdragon X60 
CloudAIR
Samsung 
MediaTeck DSS 1000
Dimensity
Nokia Dynamic Spectrum Sharing (DSS)
ZTE  SuperDSS/Magic Radio Pro 
QualComm Snapdragon X60
Samsung DSS
The Dynamic Spectrum Sharing (DSS) solution allows mobile network operators to
ZTE SuperDSS/Magic Radio Pro
offer LTE and NR services using the same frequency bands, although in an interleaved
mode.  This 
The allows 
Dynamic NR  services 
Spectrum Sharing without  the  need 
(DSS) solution of  acquiring 
allows new  and 
mobile network dedicated 
operators to fre‐
quency  spectrum,  antenna,  or  radio  frequency  units.  The  solution  is  intended  to  assist
offer LTE and NR services using the same frequency bands, although in an interleaved
mode. This allows NR services without the need of acquiring new and dedicated frequency
operators in the short‐term rollout deployment of 5G services through LTE already‐in‐use
spectrum, antenna, or radio frequency units. The solution is intended to assist operators in
spectrum. It is not intended to provide substantial performance, as that would necessitate
the short-term rollout deployment of 5G services through LTE already-in-use spectrum. It is
new dedicated spectrum for NR, but to provide coverage, reduce costs, and improve spec‐
not intended to provide substantial performance, as that would necessitate new dedicated
trum efficiency for the operator. Figure 1 presents the DSS technology with LTE and NR
spectrum for NR, but to provide coverage, reduce costs, and improve spectrum efficiency
sharing  the  same 
for the operator. frequency 
Figure band 
1 presents in  comparison 
the DSS to  using 
technology with two 
LTE and NRseparate  bands 
sharing the samefor  each
technology. 
frequency band in comparison to using two separate bands for each technology.

 
Figure 1. DSS vs. using dedicated bands for NR and LTE [31].
Figure 1. DSS vs. using dedicated bands for NR and LTE [31]. 

The deploying of DSS technology is divided into two phases: Phase 1, which is based
The deploying of DSS technology is divided into two phases: Phase 1, which is based
only on the NSA architecture and accepts a sharing ratio between 20 and 60% with a fixed
only on the NSA architecture and accepts a sharing ratio between 20 and 60% with a fixed
UL sharing ratio; and Phase 2, which introduces a dynamic UL sharing ratio and accepts
UL sharing ratio; and Phase 2, which introduces a dynamic UL sharing ratio and accepts
both NSA and SA architectures. The main differences between both phases are presented in
both NSA and SA architectures. The main differences between both phases are presented
Table 3. The sharing ratio refers to the ratio of shared resources between both technologies.
in Table 3. The sharing ratio refers to the ratio of shared resources between both technol‐
For example, considering a sharing ratio of 20% refers to 5G NR occupying 20% of the
available resources, while LTE occupies 80%. Another example is for a sharing ratio of 60%,
ogies. For example, considering a sharing ratio of 20% refers to 5G NR occupying 20% of
meaning that 5G NR uses 60% of the available resources while LTE uses only 40%.
the available resources, while LTE occupies 80%. Another example is for a sharing ratio
For downlink, the allocation of the subframes is based on Time Division Multiplexing
of 60%, meaning that 5G NR uses 60% of the available resources while LTE uses only 40%.
(TDM). In one frame, regardless of the sharing ratio adopted, subframes 0, 5, and 9 are
   
strictly dedicated to LTE transmission. Subframes 1, 2, 3, 4, 6, 7, and 8 can be used for
both LTE and NR transmission, depending on the sharing ratio and the architecture mode
adopted, see Figure 2 [32].

 
 Phase 1  Phase 2 
Sharing ratio 20–60%  Sharing ratio 5–70% 
Table 3. Comparisons between two phases of DSS implementation. 
Fixed UL sharing ratio    Dynamic UL sharing ratio 
Phase 1  Phase 2 
Sensors 2021, 21, 4215 Only NSA architecture  NSA and SA architectures 
4 of 15
Sharing ratio 20–60%  Sharing ratio 5–70% 
Supports only FDD 
Fixed UL sharing ratio   
Supports FDD and TDD 
Dynamic UL sharing ratio 
Sharing ratio update is slow 
Only NSA architecture  Sharing ratio update takes up to 10
NSA and SA architectures 
Supports only FDD 
Table 3. Comparisons Supports FDD and TDD 
between two phases of DSS implementation.
For downlink, the allocation of the subframes is based on Time Division Multip
Sharing ratio update is slow 
Phase 1
Sharing ratio update takes up to 100 ms 
Phase 2
(TDM). In one frame, regardless of the sharing ratio adopted, subframes 0, 5, and
Sharing ratio 20–60% Sharing ratio 5–70%
For downlink, the allocation of the subframes is based on Time Division Multiplexing 
strictly dedicated to LTE transmission. Subframes 1, 2, 3, 4, 6, 7, and 8 can be used fo
Fixed UL sharing ratio Dynamic UL sharing ratio
(TDM). In one frame, regardless of the sharing ratio adopted, subframes 0, 5, and 9 are 
LTE  and  NR  NSA architecturedepending  on  the NSA
Onlytransmission,  and SAratio 
sharing  architectures
and  the  architecture
strictly dedicated to LTE transmission. Subframes 1, 2, 3, 4, 6, 7, and 8 can be used for both 
Supports only FDD Supports FDD and TDD
adopted, see Figure 2 [32]. 
LTE  and Sharing
NR  transmission, 
ratio update depending 
is slow on  the  sharing 
Sharing ratio  and  the 
ratio update architecture 
takes mode 
up to 100 ms
adopted, see Figure 2 [32]. 

 
 
Figure 2. Frame allocation for DL Phase 2 DSS [32]. 
Figure 2. Frame allocation for DL Phase 2 DSS [32].
Figure 2. Frame allocation for DL Phase 2 DSS [32]. 
The  downlink  resource
The downlink resource  allocation,
allocation,  when
when  considering
considering  the
the  transmission
transmission  ofof  several
several 
frames, 
The varies  depending on 
downlink  resource the  sharing  ratio implemented 
allocation,  when  [33]. Different 
considering  the 
frames, varies depending on the sharing ratio implemented [33]. Different patterns are patterns are 
transmission  of  s
depicted  in  Figure 
depictedvaries 
in Figure 3,  for  the  NSA 
3, for the NSA the  architecture 
architecture mode.  It  can  be  observed  that  for  every 
frames,  depending on  sharing mode. It can be observed
ratio implemented  that for every patter
[33]. Different 
frame, subframes 0, 5, and 9 are always dedicated to LTE. In addition, for the first frame 
frame, subframes 0, 5, and 9 are always dedicated to LTE. In addition, for the first framethat  for
depicted  in  Figure  3,  for  the  NSA  architecture  mode.  It  can  be  observed 
only, slot 1 is represented with yellow (slot type B) and slot 2 with orange (slot type B*), 
only, slot 1 is represented with yellow (slot type B) and slot 2 with orange (slot type
frame, subframes 0, 5, and 9 are always dedicated to LTE. In addition, for the first
and both are used for transmitting synchronization and CSI‐RS signals, respectively. Ad‐
B*), and both are used for transmitting synchronization and CSI-RS signals, respectively.
only, slot 1 is represented with yellow (slot type B) and slot 2 with orange (slot ty
ditional synchronization signals are sent with a period of 20 ms in slot 1 of the remaining 
Additional synchronization signals are sent with a period of 20 ms in slot 1 of the remaining
frames, represented by green (slot type B**). The remaining slots are used for LTE and NR 
and both are used for transmitting synchronization and CSI‐RS signals, respective
frames, represented by green (slot type B**). The remaining slots are used for LTE and
transmission. 
NR transmission.
ditional synchronization signals are sent with a period of 20 ms in slot 1 of the rem
frames, represented by green (slot type B**). The remaining slots are used for LTE a
transmission. 

 
Figure 3. Resource allocation patterns for DL Phase 2 DSS [34].
Figure 3. Resource allocation patterns for DL Phase 2 DSS [34]. 

Figure 3. Resource allocation patterns for DL Phase 2 DSS [34]. 
Sensors 2021, 21, 4215
Sensors 2021, 21, 4215  5 of16 
5  of  15
 

For uplink, the allocation of the resources is based on Frequency Division Multiplex‐
For uplink, the allocation of the resources is based on Frequency Division Multiplexing
ing (FDM). As depicted in Figure 4, the Physical Resource Blocks (PRB) available for LTE 
(FDM). As depicted in Figure 4, the Physical Resource Blocks (PRB) available for LTE or
or NR transmission are represented by the color green and depend on the sharing ratio 
NR transmission are represented by the color green and depend on the sharing ratio
adopted and carrier bandwidth. Furthermore, there are seven PRBs dedicated to NR UL 
adopted and carrier bandwidth. Furthermore, there are seven PRBs dedicated to NR UL
transmission only, represented by the color yellow in the right outer edge of the frequency 
transmission only, represented by the color yellow in the right outer edge of the frequency
band. These PRBs depend on the positioning of the LTE PRACH PRBs. In Figure 4, these 
band. These PRBs depend on the positioning of the LTE PRACH PRBs. In Figure 4, these
are located in the left outer edge of the frequency band. If the LTE PRACH PRBs were 
are located in the left outer edge of the frequency band. If the LTE PRACH PRBs were
located in the right outer edge, the seven NR UL PRBs would then be positioned in the 
located in the right outer edge, the seven NR UL PRBs would then be positioned in the left
left outer edge (meaning the opposite side). 
outer edge (meaning the opposite side).

 
Figure 4. Resource allocation for UL Phase 2 DSS [34].
Figure 4. Resource allocation for UL Phase 2 DSS [34]. 
The calculation of the maximum available NR/LTE sharing ratio is given by the
The calculation of the maximum available NR/LTE sharing ratio is given by the fol‐
following equation:
lowing equation: 
 
+ PUCCHMax_LTE
N + PUCCH
N
Sharing ratio
Sharing ratio = 1 −
= 1‐
Max_LTE (1)
(1)
Max_UL
Max_UL LTE BW in PRBs  
LTE BW in PRBs
where N represents the number of PRBs that are necessary for LTE transmission, for in‐
where N represents the number of PRBs that are necessary for LTE transmission, for
stance, the LTE PRACH PRBs. 
instance, the LTE PRACH PRBs.
In this paper, we studied the concept of dynamic spectrum sharing between LTE and 
In this paper, we studied the concept of dynamic spectrum sharing between LTE and
5G NR technologies for the same mobile network operator. We assessed the performance 
5G NR technologies for the same mobile network operator. We assessed the performance
of an LTE‐NR communication system using the DSS feature, in terms of throughput, using 
of an LTE-NR communication system using the DSS feature, in terms of throughput, using
different sharing ratios for both NSA and SA architectures and for both transmission di‐
different sharing ratios for both NSA and SA architectures and for both transmission
rections (downlink and uplink). We performed a comparison of the performance while 
directions (downlink and uplink). We performed a comparison of the performance while
using different modulation schemes and numbers of layers. The remainder of the paper 
using different modulation schemes and numbers of layers. The remainder of the paper
is organized as follows. Section 2 presents the sharing ratio calculation for downlink and 
is organized as follows. Section 2 presents the sharing ratio calculation for downlink
uplink; Section 3 provides the equipment and methods used for the measurements; and 
and uplink; Section 3 provides the equipment and methods used for the measurements;
Section 4 presents the results obtained and analysis. Lastly, Section 5 delivers the conclu‐
and Section 4 presents the results obtained and analysis. Lastly, Section 5 delivers the
sions of this paper. 
conclusions of this paper.

2. Sharing Ratio Calculation

2. Sharing Ratio Calculation 
The sharing ratio between LTE and NR is defined and managed by a new system unit
The sharing ratio between LTE and NR is defined and managed by a new system unit 
denominated Common Resource Manager (CRM). It has the responsibility to compute
denominated Common Resource Manager (CRM). It has the responsibility to compute the 
the sharing ratio and update it according to traffic demands. In order to do so, the CRM
sharing ratio and update it according to traffic demands. In order to do so, the CRM con‐
continuously gathers information from both LTE and NR sites. The CRM component
tinuously gathers information from both LTE and NR sites. The CRM component is com‐
is composed by 3 objects: CRM, situated in the base station; LTE CRM, situated in the
posed by 3 objects: CRM, situated in the base station; LTE CRM, situated in the LTE sys‐
LTE
tem  system unit;
unit;  NR  NR CRM,
CRM,  situated
situated  in NR 
in  the  the NR system
system  unit.
unit.  Figure
Figure  5 presentsthe 
5  presents  themain 
main
responsibilities of the CRM. Initially, it starts by gathering information from LTE and NR

 
Sensors 2021, 21, 4215 
 

Sensors 2021, 21, 4215 6 of 15

responsibilities of the CRM. Initially, it starts by gathering information from LTE 
sites and, based on the information receives, it defines the resources to be shared
sites and, based on the information receives, it defines the resources to be shared. It then
allocates the shared resources to both technologies. According to traffic conditi
allocates the shared resources to both technologies. According to traffic conditions and
demands for a specific moment, it evaluates the optimal sharing ratio to be selec
demands for a specific moment, it evaluates the optimal sharing ratio to be selected and
finally updates it for LTE and NR. 
finally updates it for LTE and NR.

 
Figure 5. CRM main responsibilities [31].
Figure 5. CRM main responsibilities [31]. 
For downlink, depending on the traffic demands for a specific moment, the CRM
For 
receives downlink, 
information depending 
concerning load on  the  traffic 
indication demands 
and takes for ona the
a decision specific  moment,  th
DL sharing
ratio that needs to be adopted. To calculate the DL load, the weighted load
receives information concerning load indication and takes a decision on the DL  (based on PRB
occupancy), average LTE DSS Guaranteed Bit Rate (GBR) load, NR DSS GBR load, and NR
ratio that needs to be adopted. To calculate the DL load, the weighted load (based
PDCCH load need to be determined [35]. The algorithm for the sharing ratio calculation
occupancy), average LTE DSS Guaranteed Bit Rate (GBR) load, NR DSS GBR lo
for DL is presented in Figure 6. The first step consists of the verification, from the CRM
NR PDCCH load need to be determined [35]. The algorithm for the sharing ratio 
entity, of the average LTE GBR load, as well as the NR PDCCH load against a defined
tion for DL is presented in Figure 6. The first step consists of the verification, f
threshold, so that a decision can be taken regarding the resources to be assigned. If the
average LTE GBR load is higher than 70% and the NR PDCCH load is lower than 70%,
CRM entity, of the average LTE GBR load, as well as the NR PDCCH load again
then the sharing ratio for NR will decrease. Else, if the average LTE GBR load is equal or
fined threshold, so that a decision can be taken regarding the resources to be assi
lower than 70% and the NR PDCCH load is higher than 70%, then the sharing ratio for NR
the average LTE GBR load is higher than 70% and the NR PDCCH load is lower th
will increase. Lastly, if both the average LTE GBR load and NR PDCCH load are higher
then the sharing ratio for NR will decrease. Else, if the average LTE GBR load is e
than 70%, then one of the two following conditions is applied:
lower than 70% and the NR PDCCH load is higher than 70%, then the sharing r
• If the LTE GBR resource delta (n; n − 1) > 0, the sharing ratio for NR will be reduced;
•
NR  If the increase. 
will  LTE GBR resource
Lastly, delta
if  both  − 1)
(n; nthe  ≤ 0 andLTE 
average  the NR PDCCH
GBR  load resource
and  NR  delta
PDCCH  l
(n; n − 1) > 0, the sharing ratio for NR will increase.
higher than 70%, then one of the two following conditions is applied: 
The second step of the algorithm is based on the load information received from step
 TheIf the LTE GBR resource delta (n; n − 1) > 0, the sharing ratio for NR will be r
1. CRM then calculates the LTE-weighted load and the NR weighted load, from which
the LTE
If the LTE GBR resource delta (n; n − 1) ≤ 0 and the NR PDCCH resource delt
and NR total loads are determined. Finally, in step 3, the number of LTE and NR
subframes is calculated, taking into account the LTE and NR total loads from step 2. The
1) > 0, the sharing ratio for NR will increase. 
resulting number of subframes matches to a specific sharing ratio value.
The second step of the algorithm is based on the load information received fr
1. The CRM then calculates the LTE‐weighted load and the NR weighted load, from
the LTE and NR total loads are determined. Finally, in step 3, the number of LTE 
subframes is calculated, taking into account the LTE and NR total loads from ste
resulting number of subframes matches to a specific sharing ratio value. 
Sensors 2021, 21, 4215 7 of 15

 
Figure 6. Algorithm for DL Phase 2 DSS sharing ratio calculation [34].

For uplink, a similar procedure as for downlink is taken. Depending on the traffic
demands for a specific moment, the CRM receives information concerning load indication
and takes a decision on the UL sharing ratio that needs to be adopted. To calculate the UL
load, the weighted load (based on PRB occupancy) and the average LTE DSS GBR load
need to be determined. Figure 7 presents the algorithm for the UL sharing ratio calculation.
For step 1, the average LTE GBR load is verified by the CRM, so a decision can be taken
regarding the assignment of resources. If the average LTE GBR load is higher than 70%,
then the sharing ratio for NR will be decreased. The second step of the algorithm is based
on the load information received from step 1. The CRM then calculates the LTE weighted
load and the NR weighted load, from which the LTE and NR total loads are determined.
Finally, in step 3, the number of LTE and NR subframes is calculated, taking into account
the LTE and NR total loads from step 2. The resulting number of subframes matches to a
specific sharing ratio value.
1 
 
Sensors 2021, 21, 4215 8 of 15
Sensors 2021, 21, 4215  8  of  16 
 

 
Figure 7. Algorithm for UL Phase 2 DSS sharing ratio calculation [34].
Figure 7. Algorithm for UL Phase 2 DSS sharing ratio calculation [34]. 
3. Equipment and Methods
3. Equipment and Methods 
This section presents the parameters adopted for our work, as well as the scenarios
tested.This section presents the parameters adopted for our work, as well as the scenarios 
We considered a MIMO system, composed of one base station for LTE, one for
tested. We considered a MIMO system, composed of one base station for LTE, one for NR, 
NR, and one mobile station that is composed of a Qualcomm chipset prototype that is
and  one  mobile 
frequently used on station  that  is Samsung
commercial composed devices.
of  a  Qualcomm 
We used chipset  prototype 
both 64QAM and that  is  fre‐
256QAM
quently used on commercial Samsung devices. We used both 64QAM and 256QAM mod‐
modulation for the measurements. A bandwidth of 15 MHz was selected. The Absolute
ulation for the measurements. A bandwidth of 15 MHz was selected. The Absolute Radio‐
Radio-Frequency Channel Numbers (ARFCN) for NR were 175,800 for downlink and
Frequency Channel Numbers (ARFCN) for NR were 175,800 for downlink and 166,800 for 
166,800 for uplink. The NR-ARFCN is a code that refers to the carrier frequency to be used
uplink. The NR‐ARFCN is a code that refers to the carrier frequency to be used for both 
for both transmission directions of the radio channel and is defined in the 3GPP TS 38.104
transmission directions of the radio channel and is defined in the 3GPP TS 38.104 Release 
Release 16 specification [36]. The NR-ARFCN can be converted to frequency, resulting in
16 specification [36]. The NR‐ARFCN can be converted to frequency, resulting in 75,800 = 
75,800 = 879 MHz for downlink and 166,800 = 834 MHz for uplink. Frequency Division
879  MHz 
Duplex for downlink 
(FDD) was selected and 
for166,800  =  834 MHz 
all cases. for uplink. 
We performed Frequency 
throughput Division  Duplex 
measurements using
(FDD) was selected for all cases. We performed throughput measurements using physical 
physical and static equipment from Nokia Networks R&D laboratory, considering a Signal-
and static equipment from Nokia Networks R&D laboratory, considering a Signal‐to‐In‐
to-Interference-plus-Noise Ratio (SINR) higher than 25 dB and a Reference Signal Receive
terference‐plus‐Noise 
Power (RSRP) higher than Ratio 
−70(SINR) 
dBm higher  than 
with Line of 25  dB (LoS)
Sight and  a 
andReference 
withoutSignal  Receive 
the presence
ofPower (RSRP) higher than −70 dBm with Line of Sight (LoS) and without the presence of 
fading. These are standard values used at the laboratory for testing the performance
offading. These are standard values used at the laboratory for testing the performance of 
new technologies. They are considered very good radio conditions, and the reason for
new  technologies. 
choosing them is to They 
createare  considered 
almost very conditions
ideal radio good  radio inconditions,  and  the 
order to verify andreason 
confirm for 
choosing them is to create almost ideal radio conditions in order to verify and confirm the 
the aptness of DSS technology and its peak performance using physical measurements,
asaptness of DSS technology and its peak performance using physical measurements, as it 
it is a technology under development and testing. We considered both NSA and SA
is a technology under development and testing. We considered both NSA and SA archi‐
architectures. For NSA, we measured using sharing ratio values between 20 and 70%.
For SA, we measured using sharing ratio values between 30 and 70%. Table 4 below
tectures. For NSA, we measured using sharing ratio values between 20 and 70%. For SA, 
summarizes the scenario parameters adopted for the work.
we measured using sharing ratio values between 30 and 70%. Table 4 below summarizes 
the scenario parameters adopted for the work. 

 
Sensors 2021, 21, 4215  9  of  16 
 
Sensors 2021, 21, 4215 9 of 15

Table 4. Set of parameters adopted for our work. 
Table 4. Set of parameters adopted for our work.
Case  Direction  Bandwidth  Nr. of Layers  Modulation  MIMO 
1 
Case DL 
Direction 15 
Bandwidth 2  Layers
Nr. of 64QAM 
Modulation 2 × 2 
MIMO
2 1 DL DL 15 15 2 2 256QAM 
64QAM 22 × 2 
×2
3 2 DL DL 15 15 4 2 256QAM
64QAM  ×2
24 × 4 
3 DL 15 4 64QAM 4×4
4 4 DL DL 15 15 4 4 256QAM 
256QAM 44 × 4 
×4
5 5 UL UL 15 15 1 1 64QAM 
64QAM ×1
11 × 1 

Each network architecture type has different possible variations. The NSA architec‐
Each network architecture type has different possible variations. The NSA architecture
ture  is  based 
is based on theon 
LTEthe core
LTE network
core  network  and 
and uses uses  LTE‐based 
LTE-based interfaces. 
interfaces. For this For 
typethis  type  of 
of architec-
architecture, the gNodeB needs to support these interfaces and acts as a secondary node, 
ture, the gNodeB needs to support these interfaces and acts as a secondary node, while the
while the eNodeB acts as a primary or master node. There are different options to deploy 
eNodeB acts as a primary or master node. There are different options to deploy an NSA
an NSA architecture—option 3, 3a, 3x, 4, 4a, 7, and 7a. The option used for our measure‐
architecture—option 3, 3a, 3x, 4, 4a, 7, and 7a. The option used for our measurements is
ments is NSA option 3x, where the control plane is routed through the master eNodeB 
NSA option 3x, where the control plane is routed through the master eNodeB and the user
and the user plane is directly routed through the secondary gNodeB. The eNodeB also 
plane is directly routed through the secondary gNodeB. The eNodeB also communicates
communicates directly with the gNodeB and both communicate directly with the Evolved 
directly with the gNodeB and both communicate directly with the Evolved Packet Core
Packet Core (EPC). The SA architecture has 2 options—option 2 and 5. Option 2 is that 
(EPC). The SA architecture has 2 options—option 2 and 5. Option 2 is that adopted for
adopted for our work and, as it can be seen in Figure 8b, consists of a Next Generation 
our work and, as it can be seen in Figure 8b, consists of a Next Generation Core (NGC)
Core (NGC) and a gNodeB that communicates directly with it, without needing any sup‐
and a gNodeB that communicates directly with it, without needing any support of LTE
structures. For both network architectures studied, we used two radio modules that have
port of LTE structures. For both network architectures studied, we used two radio mod‐
attached to them one attenuator, as the measurements were performed in a laboratory
ules that have attached to them one attenuator, as the measurements were performed in 
with close proximity to the mobile user. We used either 2 or 4 antennas, depending on the
a laboratory with close proximity to the mobile user. We used either 2 or 4 antennas, de‐
case studied.
pending on the case studied. 

   
(a)  (b) 
Figure 8. Architecture diagram adopted for the measurements: (a) NSA architecture; (b) SA archi‐
Figure 8. Architecture diagram adopted for the measurements: (a) NSA architecture; (b) SA architecture.
tecture. 

 
Sensors 2021, 21, 4215 10 of 15

4. Results and Discussion

This section presents the results obtained from our measurements. We divide it into
two subsections: downlink and uplink results. In each subsection, we present results for
both NSA and SA architectures. We considered sharing ratios ranging from 20 to 70% for the
NSA architecture and from 30 to 70% for the SA architecture, meaning that 5G NR occupies
Sensors 2021, 21, 4215  10  of  16 
  between 20 and 70%, and 30 and 70% of the available resources, while LTE occupies between
80 and 30%, and 70 and 30% for the NSA and SA architectures, respectively.

4.1. Downlink
4. Results and Discussion 
Figure 9 presents the throughput results using the DSS feature from the first four cases
This section presents the results obtained from our measurements. We divide it into 
oftwo subsections: downlink and uplink results. In each subsection, we present results for 
Table 2, comprising the NSA architecture. The differences between the cases consist of
the modulation type, that is either 64QAM or 256QAM modulation, and MIMO type, that
both NSA and SA architectures. We considered sharing ratios ranging from 20 to 70% for 
either 2 × 2 or 4 × 4 MIMO. Each case depicts five different curves. The green (NR only)
is the NSA architecture and from 30 to 70% for the SA architecture, meaning that 5G NR 
and purple (LTE only) curves represent the values for the throughput achieved without the
occupies between 20 and 70%, and 30 and 70% of the available resources, while LTE oc‐
use of DSS technology. The yellow curve (DSS LTE + NR) is the most important one as it
cupies between 80 and 30%, and 70 and 30% for the NSA and SA architectures, respec‐
presents
tively.  the total throughput obtained with DSS. The remaining blue (DSS LTE) and red
(DSS NR) curves represent the individual throughputs for each technology while using
DSS. Notice that the DSS LTE + NR throughput equals the sum of the individual DSS
4.1. Downlink 
LTE and DSS NR throughputs. It can be observed that, for all cases, when we increased
Figure  9  presents  the  throughput  results  using  the  DSS  feature  from  the  first  four 
the sharing ratio, the values for the DSS NR TP also increased while the DSS LTE TP
cases of Table 2, comprising the NSA architecture. The differences between the cases con‐
values decreased. This was an expected behavior, as the higher the sharing ratio, the more
sist  of  the  modulation  type,  that is  either  64QAM  or  256QAM  modulation,  and  MIMO 
resources will be available for NR transmission and the fewer for LTE transmission. It can
type, that is either 2 × 2 or 4 × 4 MIMO. Each case depicts five different curves. The green 
also be observed that from a sharing ratio of approximately 57%, meaning that NR occupies
57%(NR only) and purple (LTE only) curves represent the values for the throughput achieved 
of the resources while LTE occupies 43%, the individual throughput for NR surpassed
thewithout  the  use  of  In
LTE throughput. DSS  technology. 
addition, The  yellow 
it is visible curve 
that the (DSS 
overall LTE  +  NR) 
throughput is  the 
using DSSmost 
was im‐
portant one as it presents the total throughput obtained with DSS. The remaining blue 
slightly lower than that for the LTE or NR-only throughputs. This is understandable, as the
(DSS LTE) and red (DSS NR) curves represent the individual throughputs for each tech‐
available frame resources are shared between both technologies.
nology while using DSS. Notice that the DSS LTE + NR throughput equals the sum of the 
Comparing case 1 and case 2, where the difference is the modulation type that increases
individual DSS LTE and DSS NR throughputs. It can be observed that, for all cases, when 
from 64QAM to 256QAM modulation, it can be concluded that the major difference is on
we increased the sharing ratio, the values for the DSS NR TP also increased while the DSS 
the maximum values for throughput. For case 1, depending on the sharing ratio adopted,
LTE TP values decreased. This was an expected behavior, as the higher the sharing ratio, 
between 90 and 100 Mbps were obtained, while for case 2, the values for throughput were
the more resources will be available for NR transmission and the fewer for LTE transmis‐
approximately 120–135 Mbps. An increase of 35% was observed. A similar comparison
cansion. It can also be observed that from a sharing ratio of approximately 57%, meaning that 
be conducted for cases 3 and 4. For case 3, the maximum DSS throughput values were
NR occupies 57% of the resources while LTE occupies 43%, the individual throughput for 
175–200 Mbps, while for case 4, the values varied between 240 and 260 Mbps, depending
onNR surpassed the LTE throughput. In addition, it is visible that the overall throughput 
the sharing ratio. For these cases, an increase of 37% in throughput was observed for a
using DSS was slightly lower than that for the LTE or NR‐only throughputs. This is un‐
sharing ratio of 20%, while for a 70% sharing ratio, the throughput increase was 30% (see
derstandable, as the available frame resources are shared between both technologies. 
Table 5).

   
(a)  (b) 
Figure 9. Cont.
Sensors 2021, 21, 4215
Sensors 2021, 21, 4215  1111 
ofof 
15 16 
 

 
(c)  (d) 
Figure 9. DL throughput results for NSA architecture, with sharing ratios from 20% up to 70%: (a) Case 1 results; (b) Case 
Figure 9. DL throughput results for NSA architecture, with sharing ratios from 20% up to 70%: (a) Case 1 results; (b) Case 2
2 results; (c) Case 3 results; (d) Case 4 results. 
results; (c) Case 3 results; (d) Case 4 results.

TableComparing case 1 and case 2, where the difference is the modulation type that in‐
5. DL throughput results for DSS LTE + NR in Mbps.
creases from 64QAM to 256QAM modulation, it can be concluded that the major differ‐
ence is on the maximum values for throughput. For case 1, depending on the sharing ratio 
DL NSA DSS LTE + NR (Mbps) DL SA DSS LTE + NR (Mbps)
30% 40% adopted,  50% between  60% 90  and  100  Mbps 
70% 30%were  obtained, 
40% while 
50% for  case 
60%2,  the  values 
70% for 
Case 1 97.5 95.9 throughput were approximately 120–135 Mbps. An increase of 35% was observed. A sim‐
94.4 92.9 91.4 90.9 88.3 85.8 83.2 80.7
Case 2 132.9 130.6 ilar comparison can be conducted for cases 3 and 4. For case 3, the maximum DSS through‐
128.3 126 123.8 123.9 120.1 116.4 112.7 109
Case 3 192.6 188.8 put values were 175–200 Mbps, while for case 4, the values varied between 240 and 260 
185 181.2 177.4 179.8 174 168.1 162.3 156.5
Case 4 257.1 251.4 245.7 240 234 239.9 231.3 222.7 214.2 205.6
Mbps, depending on the sharing ratio. For these cases, an increase of 37% in throughput 
was observed for a sharing ratio of 20%, while for a 70% sharing ratio, the throughput 
Regarding case 1 and case 3, where the difference consists of the number of transmit-
increase was 30% (see Table 5). 
ting and receiving antennas, 2 × 2 MIMO and 4 × 4 MIMO, respectively, an increase in
allTable 5. DL throughput results for DSS LTE + NR in Mbps. 
throughput values of approximately 50% could be observed, along with an increase
in complexity.
 
Figure 10DL NSA DSS LTE + NR (Mbps) 
depicts the DL throughput results with the DL SA DSS LTE + NR (Mbps) 
DSS feature for the first four cases
from Table
  2,30% 
using the
40% SA architecture,
50%  60%  contrary
70% to the results40% 
30%  from Figure
50%  8 that
60% made use
70% 
of the NSA architecture. The main difference of both types of architectures is that NSA is
Case 1  97.5  95.9  94.4  92.9  91.4  90.9  88.3  85.8  83.2  80.7 
an intermediary solution that is based on the LTE network, while the SA architecture does
notCase 2 
depend132.9  130.6 on the
in any way 128.3  126  123.8 
LTE network, using a123.9  120.1  116.4 
Next-Generation Core112.7  109 
(NGC) along
with NR protocols.
Case 3  192.6  Moreover,
188.8  185  the SA181.2 
architecture
177.4  leads to an 174 
179.8  improved efficiency
168.1  162.3 with less
156.5 
complexity. Comparing case 1 and case 2, we can observe that maximum DSS throughput
Case 4 
values 257.1 
varied 251.4 
between 245.7 
80 and 240 and 234 
90 Mbps, 110 and239.9  231.3 
120 Mbps, 222.7  214.2 
respectively. 205.6 
The increase
from one case to another was approximately 36%. For cases 3 and 4, the DSS values varied
between Regarding case 1 and case 3, where the difference consists of the number of transmit‐
160 and 180 Mbps, and 200 and 240 Mbps, respectively. For these, an increase of
ting and receiving antennas, 2 × 2 MIMO and 4 × 4 MIMO, respectively, an increase in all 
approximately 29% was observed.
throughput 
Regarding values 
case 2of 
andapproximately 50% 
case 4, the increasecould 
in thebe  observed, along 
DSS throughput was with an increase 
approximatelyin 
complexity. 
87% for both sharing ratios of 30% and 70%. In addition, it can be remarked that for all
Figure 10 depicts the DL throughput results with the DSS feature for the first four 
cases, the values for the NR-only throughput for the SA architecture were smaller than
cases from Table 2, using the SA architecture, contrary to the results from Figure 8 that 
those of the NSA architecture, with a difference of around 15 Mbps for case 1, 20 Mbps for
made use of the NSA architecture. The main difference of both types of architectures is 
cases 2 and 3, and 40 Mbps for case 4. The reason for this is that in the SA architecture, the
that NSA is an intermediary solution that is based on the LTE network, while the SA ar‐
number of broadcast signals was higher than that in the NSA architecture. An example is
chitecture does not depend in any way on the LTE network, using a Next‐Generation Core 
the presence of System Information Block (SIB) signals, as well as paging with information
regarding the cell.
(NGC) along with NR protocols. Moreover, the SA architecture leads to an improved ef‐
ficiency with less complexity. Comparing case 1 and case 2, we can observe that maximum 
DSS throughput values varied between 80 and 90 Mbps, and 110 and 120 Mbps, respec‐
tively. The increase from one case to another was approximately 36%. For cases 3 and 4, 
the DSS values varied between 160 and 180 Mbps, and 200 and 240 Mbps, respectively. 
For these, an increase of approximately 29% was observed. 

 
Sensors 2021, 21, 4215 
Sensors 2021, 21, 4215 12 of 1512  of  16 
 

   
(a)  (b) 

 
(c)  (d) 
Figure 10. DL throughput results for SA architecture, with sharing ratios from 30% up to 70%: (a) Case 1 results; (b) Case 
Figure 10. DL throughput results for SA architecture, with sharing ratios from 30% up to 70%: (a) Case 1 results; (b) Case 2
2 results; (c) Case 3 results; (d) Case 4 results. 
results; (c) Case 3 results; (d) Case 4 results.

Table 6 provides the percentage loss of the throughput that occurs when using DSS
Regarding case 2 and case 4, the increase in the DSS throughput was approximately 
instead of an NR-only system. It can be observed that there was a loss in throughput values
87% for both sharing ratios of 30% and 70%. In addition, it can be remarked that for all 
between a minimum of 10% and a maximum of 26%, depending on the sharing ratio and
cases, the values for the NR‐only throughput for the SA architecture were smaller than 
case adopted. The existence of a decrease in throughput was expected, as with DSS, the
those of the NSA architecture, with a difference of around 15 Mbps for case 1, 20 Mbps for 
available resources are shared with LTE, and hence, there are fewer resources available for
cases 2 and 3, and 40 Mbps for case 4. The reason for this is that in the SA architecture, the 
NR, compared to a system that is based only on NR. However, from the results obtained,
number of broadcast signals was higher than that in the NSA architecture. An example is 
for the NSA architecture, having a loss between 14 and 25% is not a considerable decrease
the presence of System Information Block (SIB) signals, as well as paging with information 
taking in account the fact that there is no need for new dedicated spectrum to be allocated
for NR, as it shares the bandwidth with LTE technology. For case 1, the average loss was
regarding the cell. 
19.8% for NSA and 17.2% for SA. For case 2, the loss was 17.5% for NSA and 15.6% for SA.
Table 6 provides the percentage loss of the throughput that occurs when using DSS 
For case 3, we had a 21.2% loss for NSA and 19% for SA. Lastly, for case 4, the loss was
instead of an NR‐only system. It can be observed that there was a loss in throughput val‐
20.6% for NSA and 19.2% for SA.
ues between a minimum of 10% and a maximum of 26%, depending on the sharing ratio 
We can observe that the decrease in the DSS throughput was higher with the increase
and case adopted. The existence of a decrease in throughput was expected, as with DSS, 
in the sharing ratio. This is due to the fact that, as the sharing ratio increases (meaning that
the available resources are shared with LTE, and hence, there are fewer resources availa‐
a higher number of the available resources were used for NR transmission and fewer for
ble for NR, compared to a system that is based only on NR. However, from the results 
LTE), more synchronization signals are transmitted in the slots dedicated to NR, as well as
obtained, for the NSA architecture, having a loss between 14 and 25% is not a considerable 
overhead signals.
decrease taking in account the fact that there is no need for new dedicated spectrum to be 
allocated for NR, as it shares the bandwidth with LTE technology. For case 1, the average 
loss was 19.8% for NSA and 17.2% for SA. For case 2, the loss was 17.5% for NSA and 
15.6% for SA. For case 3, we had a 21.2% loss for NSA and 19% for SA. Lastly, for case 4, 
the loss was 20.6% for NSA and 19.2% for SA. 
   
 Sharing 
20%  30%  40%  50%  60%  70%  30%  40%  50%  60%  70% 
Ratio 
  NSA ARCHITECTURE  SA ARCHITECTURE 
Case 1  −17%  −18%  −19%  −20%  −22%  −23%  −12%  −15%  −17%  −20%  −22% 
Sensors 2021, 21, 4215 13 of 15
CASE 2  −14%  −15%  −17%  −18%  −20%  −21%  −10%  −13%  −16%  −18%  −21% 
CASE 3  −17%  −19%  −20%  −22%  −24%  −25%  −13%  −16%  −19%  −22%  −25% 
Table 6. LossCASE 4  −16%  −18% 
of DL throughput, −20% 
in %, of −22% 
the DSS LTE +−23%  −25% 
NR value −13%  −16% 
in comparison −19%  −22%  −26% 
to NR only.

Sharing Ratio 20% 30% 40% 50% 60% 70% 30% 40% 50% 60% 70%
We can observe that the decrease in the DSS throughput was higher with the increase 
NSA ARCHITECTURE SA ARCHITECTURE
in the sharing ratio. This is due to the fact that, as the sharing ratio increases (meaning that 
Case 1 −17% −a higher number of the available resources were used for NR transmission and fewer for 
18% −19% −20% −22% −23% −12% −15% −17% −20% −22%
Case 2 −14% −15% −17% −18% −20% −21% −10% −13% −16% −18% −21%
Case 3 −17% −LTE), more synchronization signals are transmitted in the slots dedicated to NR, as well 
19% −20% −22% −24% −25% −13% −16% −19% −22% −25%
Case 4 −16% −as overhead signals. 
18% −20% −22% −23% −25% −13% −16% −19% −22% −26%

4.2. Uplink 
4.2. Uplink
Figure 11 depicts the UL throughput results for case 5 from Table 6, using both NSA 
Figure 11 depicts the UL throughput results for case 5 from Table 6, using both NSA
and SA architecture. Regarding the NSA architecture, it can be observed that the maxi‐
and SA architecture. Regarding the NSA architecture, it can be observed that the maximum
mum throughput values for LTE and NR only (not considering the DSS feature) were 40 
throughput values for LTE and NR only (not considering the DSS feature) were 40 and
and 22 Mbps, respectively. Moreover, if we compare the results of the DSS LTE + NR and 
22 Mbps, respectively. Moreover, if we compare the results of the DSS LTE + NR and
NR‐only throughputs from both architecture types, we can conclude that they were simi‐
NR-only throughputs from both architecture types, we can conclude that they were similar.
lar. Therefore, there is no difference between the SA and NSA architecture, as there are no 
Therefore, there is no difference between the SA and NSA architecture, as there are no
additional channels that need to be transmitted and hence occupy extra resources. 
additional channels that need to be transmitted and hence occupy extra resources.

 
(a)  (b) 
Figure 11. UL throughput results for case 5. (a) NSA architecture; (b) SA architecture. 
Figure 11. UL throughput results for case 5. (a) NSA architecture; (b) SA architecture.

Table 7 presents the percentage loss of throughput when using the DSS technology.
Table 7 presents the percentage loss of throughput when using the DSS technology. 
We can observe that for both architecture types, a maximum loss of 25% occurred for a
We can observe that for both architecture types, a maximum loss of 25% occurred for a 
sharing ratio of 70%, meaning that NR occupied 70% of the available resources while LTE
sharing ratio of 70%, meaning that NR occupied 70% of the available resources while LTE 
occupied only 30%. Therefore, we can conclude that using the DSS technology comes with
occupied only 30%. Therefore, we can conclude that using the DSS technology comes with 
a compromise of a maximum 25% loss on throughput. However, the decrease observed is
a compromise of a maximum 25% loss on throughput. However, the decrease observed is 
not that considerable if taking into account that, instead of needing an additional 15 MHz of
not that considerable if taking into account that, instead of needing an additional 15 MHz 
of bandwidth,
bandwidth, the system shares thethe 
the system shares  actual 15 MHz
actual 15  between both both 
MHz between  technologies. Subsequently,
technologies.  Subse‐
from the operator’s point of view, DSS brings the advantage of cost reduction and
quently, from the operator’s point of view, DSS brings the advantage of cost reduction  spectrum
efficiency, while being able to present the “5G icon game” and provide coverage strategies.
and spectrum efficiency, while being able to present the “5G icon game” and provide cov‐
erage strategies. 
Table 7. Loss of UL throughput, in %, of the DSS LTE + NR value in comparison to NR only.
   
Sharing Ratio 20% 30% 40% 50% 60% 70% 30% 40% 50% 60% 70%
NSA ARCHITECTURE SA ARCHITECTURE
Case 5 −2% −5% −10% −17% −20% −25% −5% −10% −17% −20% −25%
 
5. Conclusions
In this paper, we analyzed the impact and the advantages of using the DSS technology
in an LTE-NR communication system. We proposed different schemes for the resource
allocation, according to the selected sharing ratio. The results obtained provided insight
into the behavior of the system with DSS and showed that it is a technology that brings
advantages from the operator’s point of view, mainly regarding the spectrum efficiency
Sensors 2021, 21, 4215 14 of 15

and cost reduction. In conclusion, from the results obtained, it is clear that using the DSS
technology brings a major advantage of not needing extra dedicated bandwidth for NR
systems, which, from the operator’s point of view, leads to an improvement of spectrum
efficiency and cost reduction. We performed a comparison of the spectrum usage for
LTE and NR when adopting the DSS feature and without. In addition, we measured the
throughput obtained for both LTE and NR, using the proposed allocation schemes for each
sharing ratio. The results obtained clearly demonstrated that even if a maximum loss of
25% on throughput is observed, there is a major advantage in using the DSS technology due
to the fact that there is a cost reduction for the mobile operator alongside an optimization
on the spectrum usage, due to the fact that the MNO can re-use the already existing 15 MHz
bandwidth of LTE and does not need to buy any new dedicated supplementary 15 MHz
for 5G services.
In conclusion, the deployment of DSS technology is useful, especially for the initial
deployment of 5G NR, as the operator is able to build strategies while presenting the initial
5G picture to the consumer, through LTE already-in-use spectrum.

Author Contributions: Conceptualization, G.B., F.A. and M.O.; software, G.B.; validation, G.B., F.A.
and M.O.; formal analysis, G.B., F.A. and M.O.; investigation, G.B. and F.A.; resources, G.B. and F.A.;
data curation, G.B. and F.A.; writing—original draft preparation, G.B.; writing—review and editing,
G.B. and F.A.; project administration, F.A.; funding acquisition, F.A. All authors have read and agreed
to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Acknowledgments: This work was developed using Nokia’s equipment and infrastructure. For this
reason, the authors wish to acknowledge the support provided by Nokia.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design
of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or
in the decision to publish the results.

References
1. Shafi, M.; Molisch, A.F.; Smith, P.J.; Haustein, T.; Zhu, P.; De Silva, P.; Tufvesson, F.; Benjebbour, A.; Wunder, G. 5G: A Tutorial
Overview of Standards, Trials, Challenges, Deployment, and Practice. IEEE J. Sel. Areas Commun. 2017, 35, 1201–1221. [CrossRef]
2. Adebusola, J.A.; Ariyo, A.A.; Elisha, O.A.; Olubunmi, A.M.; Julius, O.O. An Overview of 5G Technology. In Proceedings of the
2020 International Conference in Mathematics, Computer Engineering and Computer Science (ICMCECS), Ayobo, Nigeria, 18–21
March 2020; pp. 1–4.
3. Rhayour, A.E.; Mazri, T. 5G Architecture: Deployment scenarios and options. In Proceedings of the 2019 International Symposium
on Advanced Electrical and Communication Technologies (ISAECT), Rome, Italy, 27–29 November 2019; pp. 1–6.
4. Agiwal, M.; Roy, A.; Saxena, N. Next generation 5G wireless networks: A comprehensive survey. IEEE Commun. Surv. Tutor. 2016,
18, 1617–1655. [CrossRef]
5. Dymkova, S.S. Breakthrough 5G data call using dynamic spectrum sharing to accelerate nationwide 5G deployments. Synchroinfo
J. 2019, 5, 17–21.
6. 5G NR and 4G LTE Coexistence: A Comprehensive Deployment Guide to Dynamic Spectrum Sharing [Online]. Available
online: https://newsletter.mediatek.com/hubfs/mediatek5gprogress/Dynamic-Spectrum-Sharing-WhitePaper-PDFDSSWP-
031320.pdf (accessed on 27 May 2021).
7. Tikhvinskiy, V.; Deviatkin, E.; Aitmagambetov, A.; Kulakaeva, A. Provision of IoT Services for CO-Located 4G/5G Networks
Utilization with Dynamic Frequency Sharing. In Proceedings of the 2020 International Conference on Engineering Management
of Communication and Technology (EMCTECH), Vienna, Austria, 20–22 October 2020; pp. 1–4. [CrossRef]
8. Number of IoT Connected Devices Worldwide in 2018, 2025 and 2030 [Online]. Available online: https://www.statista.com/
statistics/802690/worldwide-connected-devices-by-access-technology/ (accessed on 29 May 2021).
9. Pandit, S.; Singh, G. An overview of spectrum sharing techniques in cognitive radio communication system. Wirel. Netw. 2017,
23, 497–518. [CrossRef]
10. Matinmikko-Blue, M.; Yrjölä, S.; Seppänen, V.; Ahokangas, P.; Hämmäinen, H.; Latva-aho, M. Analysis of Spectrum Valuation
Approaches: The Viewpoint of Local 5G Networks in Shared Spectrum Bands. In Proceedings of the 2018 IEEE International
Symposium on Dynamic Spectrum Access Networks (DySPAN), Seoul, Korea, 22–25 October 2018; pp. 1–9.
Sensors 2021, 21, 4215 15 of 15

11. Li, S.; Da Xu, L.; Zhao, S. The internet of things: A survey. Inf. Syst. Front. 2015, 17, 243–259. [CrossRef]
12. Ning, H.; Liu, H.; Ma, J.; Yang, L.T.; Wan, Y.; Ye, X.; Huang, R. From Internet to Smart World. IEEE Access 2015, 3, 1994–1999.
[CrossRef]
13. Chilipirea, C.; Ursache, A.; Popa, D.O.; Pop, F. Energy efficiency and robustness for IoT: Building a smart home security system.
In Proceedings of the (ICCP), Cluj-Napoca, Romania, 8–10 September 2016; pp. 43–48.
14. Tehrani, R.H.; Vahid, S.; Triantafyllopoulou, D.; Lee, H.; Moessner, K. Licensed Spectrum Sharing Schemes for Mobile Operators:
A Survey and Outlook. IEEE Commun. Surv. Tutor. 2016, 18, 2591–2623. [CrossRef]
15. Zhang, L.; Xiao, M.; Wu, G.; Alam, M.; Liang, Y.-C.; Li, S. A survey of advanced techniques for spectrum sharing in 5G networks.
IEEE Wirel. Commun. 2017, 24, 44–51. [CrossRef]
16. Gandotra, P.; Jha, R.K.; Jain, S. Green communication in next generation cellular networks: A survey. IEEE Access 2017, 5,
11727–11758. [CrossRef]
17. Massaro, M.; Beltrán, F. Will 5G lead to more spectrum sharing? Discussing recent developments of the LSA and the CBRS
spectrum sharing frameworks. Telecommun. Policy 2020, 44, 101973. [CrossRef]
18. Jorswieck, E.A.; Badia, L.; Fahldieck, T.; Karipidis, E.; Luo, J. Spectrum sharing improves the network efficiency for cellular
operators. IEEE Commun. Mag. 2014, 52, 129–136. [CrossRef]
19. Yu, G.; Li, G.Y.; Wang, L.; Maaref, A.; Lee, J.; Lopez-Perez, D. Guest Editorial: LTE in Unlicensed Spectrum. IEEE Wirel. Commun.
2016, 23, 6–7. [CrossRef]
20. Guohua, Z.; Tianle, D.; Li, Y. A dynamic spectrum re-allocation scheme in GSM and LTE co-existed networks. In Proceedings of
the 2014 International Symposium on Wireless Personal Multimedia Communications (WPMC), Sydney, NSW, Australia, 7–10
September 2014; pp. 595–600.
21. Jo, M.; Klymash, M.; Maksymyuk, T.; Kozlovskiy, R. Dynamic spectrum sharing algorithm for combined mobile networks. In
Proceedings of the 2014 20th International Conference on Microwaves, Radar and Wireless Communications (MIKON), Gdansk,
Poland, 16–18 June 2014; pp. 1–3.
22. Irnich, T.; Kronander, J.; Selén, Y.; Li, G. Spectrum sharing scenarios and resulting technical requirements for 5G systems. In
Proceedings of the 2013 IEEE 24th International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC
Workshops), London, UK, 8–9 September 2013; pp. 127–132. [CrossRef]
23. Whitmore, A.; Agarwal, A.; Xu, L.D. The Internet of Things. A survey of topics and trends. Inf. Syst. Front. 2015, 17, 261–274.
[CrossRef]
24. Palattella, M.R.; Dohler, M.; Grieco, A.; Rizzo, G.; Torsner, J.; Engel, T.; Ladid, L. Internet of Things in the 5G Era: Enablers,
Architecture, and Business Models. IEEE J. Sel. Areas Commun. 2016, 34, 510–527. [CrossRef]
25. Zhang, L.; Liang, Y.; Xiao, M. Spectrum Sharing for Internet of Things: A Survey. IEEE Wirel. Commun. 2019, 26, 132–139.
[CrossRef]
26. Lin, S.; Kong, L.; Gao, Q.; Khan, M.K.; Zhong, Z.; Jin, X.; Zeng, P. Advanced Dynamic Channel Access Strategy in Spectrum
Sharing 5G Systems. IEEE Wirel. Commun. 2017, 24, 74–80. [CrossRef]
27. Andrews, J.G.; Buzzi, S.; Choi, W.; Hanly, S.V.; Lozano, A.; Soong, A.C.; Zhang, J.C. What Will 5G Be? IEEE J. Sel. Areas Commun.
2014, 32, 1065–1082. [CrossRef]
28. Bogucka, H.; Kryszkiewicz, P.; Kliks, A. Dynamic spectrum aggregation for future 5G communications. IEEE Commun. Mag.
2015, 53, 35–43. [CrossRef]
29. Yang, Y.; Dai, L.; Li, J.; Mumtaz, S.; Rodriguez, J. Optimal Spectrum Access and Power Control of Secondary Users in Cognitive
Radio Networks. EURASIP J. Wirel. Commun. Netw. 2017, 98, 1–15. [CrossRef]
30. Hu, F.; Chen, B.; Zhu, K. Full Spectrum Sharing in Cognitive Radio Networks toward 5G: A Survey. IEEE Access 2018, 6,
15754–15776. [CrossRef]
31. Nokia Networks. LTE-NR DSS Phase I; Network Engineering: Espoo, Finland, 2020.
32. Barb, G.; Otesteanu, M.; Roman, M. Dynamic Spectrum Sharing for LTE-NR Downlink MIMO Systems. In Proceedings of the
2020 International Symposium on Electronics and Telecommunications (ISETC), Timisoara, Romania, 5–6 November 2020; pp. 1–4.
[CrossRef]
33. Bhattarai, S.; Park, J.J.; Gao, B.; Bian, K.; Lehr, W. An overview of dynamic spectrum sharing: Ongoing initiatives, challenges, and
a roadmap for future research. IEEE Trans. Cogn. Commun. Netw. 2016, 2, 110–128. [CrossRef]
34. Nokia Networks. LTE-NR DSS with CRS Rate Matching; Network Engineering: Espoo, Finland, 2020.
35. Kliks, A.; Kryszkiewicz, P. Multichannel simultaneous uplink and downlink transmission scheme for flexible duplexing. J. Wirel.
Com. Netw. 2017, 2017, 111. [CrossRef]
36. 5G; NR; Base Station (BS) Radio Transmission and Reception. 3GPP TS 38.104 Version 16.7.0 Release 16, April 2021 [Online].
Available online: https://www.etsi.org/deliver/etsi_ts/138100_138199/138104/16.07.00_60/ts_138104v160700p.pdf (accessed
on 18 June 2021).