See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321336233

Investigation of the Impact of Heterogeneous Traffic on Performance of LTE-

based Mission Critical Communication Networks

Conference Paper · November 2017

DOI: 10.1109/IEMCON.2017.8117173

CITATIONS READS

0 110

4 authors, including:

Ivica Kostanic Olawale E. Olasupo

Florida Institute of Technology Intel
77 PUBLICATIONS   1,305 CITATIONS    15 PUBLICATIONS   165 CITATIONS   

SEE PROFILE SEE PROFILE

Humaid Saif Alshamsi

Florida Institute of Technology
18 PUBLICATIONS   253 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

artificial intelligence View project

Global Positioning System View project

All content following this page was uploaded by Humaid Saif Alshamsi on 29 May 2018.

The user has requested enhancement of the downloaded file.

 Investigation of the Impact of Heterogeneous Traffic
on Performance of LTE-based Mission Critical
Communication Networks
Kehinde O. Olasupo, Kostanic I, Tajudeen O. Olasupo
Department of Electrical Engineering, and Department of Engineering
Computer Engineering, Florida Institute of and Computer Science
Technology, Seattle Pacific University
150 W. University Blvd, Melbourne, FL 32901, 3307 3rd Ave West, Seattle WA 98119,
USA, kolasupo2011@my.fit.edu
Abstract— The existing Private/Professional Mobile Radio
(PMR) technologies designed specifically for Mission Critical
Communication (MCC) systems are narrowband and wideband
devices, with limited network data capacity in emergency
scenarios. They are majorly used to support MCC voice
communications and low data rate applications during mission
critical operations. However, the need for broadband systems
that would support high radio data capacity keep increasing
during major incidents and accident scenarios. Because of this,
the MCC agencies were attracted by the broadband capabilities
of Long Term Evolution (LTE) technology. But, the capacity of
LTE- based MCC systems is still a concern. Therefore, this study
describes simulation approach to model heterogeneous data
applications over the LTE-based MCC network. The relation
between the traffic load (video, data, and voice, short messaging)
and waiting time is presented. ARENA simulation tool is used to
show the throughput, waiting time, and resource utilization to be
expected when using LTE-based MCC networks. The simulation
results are compared with the analytical and 3GPP models.
Results show that up to ten users each with traffic less than 3.6
Mbps can simultaneously upload data on the uplink of the LTE-
based MCC network. The results from this study can help the
network designers in the implementation of equipment and
devices that could support MCC services over LTE networks.
Keywords— ARENA; broadband; LTE; mission critical
communications; private mobile radio; throughput; waiting time;
I. INTRODUCTION
The PMR systems used for MCC, are conventional private
radios that enable communications between a base station and
multiple mobile radios. PMR devices are primarily used by
the public protection disaster relief (PPDR) agencies such as
fire fighters, police, border security and ambulance services.
This kind of communication systems is to ensure safety of
people in emergency scenarios everywhere in the world. The
major requirements for the PMRs to be used for MCC include:
ability to perform satisfactorily over a long period of time,
ability to function effectively under adverse conditions,
secured communication and must be able to support group
communications. The PMR technologies include Terrestrial
Trunked Radio (TETRA), TETRA for Police (TETRAPOL),
and Associated Public Safety Communication Officials
Humaid Alshamsi
Department of Electrical Engineering, and
Computer Engineering, Florida Institute
of Technology,
150 W. University Blvd, Melbourne,
USA FL 32901, USA
(APCO) Project 25. These technologies are narrowband
designed initially only for voice communications and low data
rate applications [1], [2], [3]. Though, most of these radio
technologies are presently evolving to a technology with
increased data rate applications (wideband) [4]. But, they are
not sufficient to meet the public safety agencies’ high data rate
needs. The agencies need a technology that can support
internet access— there may be need to access architectural
plans of buildings in order to locate harmful objects in an
incidence scene. Also, there should be support for possibility
of video streaming and live video feed in emergency cases.
There may be an urgent need to upload image or video to assist
remote medical team. In order to meet all these high data
services of the MCC users, LTE has been accepted by many
organizations across the world as technology for broadband
wireless mission critical communications and services [5], [6],
[7]. Among many important features that make LTE to be so
desired for MCC are: high date rate both in the downlink and
uplink, possibility of deployment with a variety of channel
bandwidth, higher spectrum efficiency and low latency.
However, there is much concerns on the network capacity of
LTE-based MCC systems. The PPDR agencies are likely to
experience a higher demand for uplink capacity because most
of the videos and data are uplink to the control center
(dispatcher). What will be the impact of different traffic- data,
voice, video and short messages on the performance of LTE
uplink in emergency situations? The aim of this research is to
tackle the challenges and propose solutions to the question
above. The paper looks at the performance of LTE-based MCC
network in terms of channel utilization, throughput, delay, and
message delivery ratio as the metrics of performance. This is
done under various traffic load, different traffic generation,
service distributions and different simulation time.
In order to achieve the aim of this study, a single eNodeB
is considered because in MCC networks, the capacity is not
distributed across wider area as it is experienced for
commercial LTE networks. Here, emergency scenarios will
normally happen in one location and one eNodeB will be
sufficient for successful data transmission. This study uses
ARENA as simulation tool to investigate the impact of
heterogeneous traffic on the performance of LTE-based MCC
networks. The simulation is based on different scenarios of
different data applications on the air interface.
The remainder of this paper is organized as follows:
Section II reviews some literature on the previous research.
Section III presents LTE-based MCC system architecture.
Section IV provides the simulation scenarios and some
assumptions taken. Simulation results are discussed in Section
V. Finally, section VI has the summary of the study and
recommends prospects for future work.
II. RELATED LITERATURE
The first process of the deployment of MCC on LTE started
in USA. The government allocated 2 x 10 MHz of 700 MHz
spectrum to public safety users on LTE band 14 [8]. This paper
reviews literature and finds very few research on the
performance of LTE- based MCC systems. The reason is
because the final specifications and requirements for mission
critical services by 3GPP is still in progress and will be
available soon. Although, a lot of work has been published on
the traffic modeling of trunked dispatch PMR systems. [9],
[10], [11], [12]. But, these technologies are not broadband that
is required for MCC presently. However, some studies have
been published on the deployment of LTE for public safety,
there is still more to be done to evaluate its performance in
terms of capacity. The authors in [13] analyze the push to talk
voice traffic over LTE network. They use blocking probability
to analyze the queue model. In [14], the authors discuss the
signaling performance of LTE-based MCC networks in terms
of the call set up time. In an effort to evaluate the network
performance, authors in [15] present the analytical approach to
demonstrate the LTE network capacity for public safety
communications; but there is still need to demonstrate further
with simulations approach. A more detailed model of the
proposed network can be emulated in computer simulations.
III. LTE–BASED MCC SYSTEM ARCHITECTURE
AND REQUIREMENTS
A. LTE Requirements for MCC
LTE is developed by the Third Generation Partnership
Project (3GPP) as a technology to support broadband
applications for longer years in to the future [16]. LTE is an all
Internet Protocol (IP) based technology that is capable of
providing high data rates up to 100 Mbps in the downlink (DL)
and 50 Mbps in the uplink (UL). There are some important
requirements that make LTE to be accepted for MCC
applications [17]:
1. High data rates and low latency that users would
require.
2. LTE is designed to work with multiple channel
Bandwidths of 1.4, 3, 5, 10, 15 or 20 MHz in both DL
and UL.
3. Higher spectrum efficiency
4. High level of security. Since the traffic generated,
stored and exchanged by the MCC users are highly
sensitive and confidential.
5. Priority and Pre-emption mechanisms are already in-
built features designed for LTE and can be readily
used for MCC.
6. Quality of Service applications in LTE allows traffic
with high requirements on throughput an latency (e.g.
video streaming) to have preference over low data
applications (short message).
Further improvements in capabilities and performance of
LTE for MCC services are described in the 3GPP releases 12
– 15 [18]. Proximity services and group communications
system enablers are described in release 12. Mission critical
push to talk (MCPTT) is completed in release 13 and expected
to be available by the end of year 2017. Also, in release 14,
there is enhancement to MCPTT, MC-Data and MC-Video
[19]. Release 14 specifications will be completed by the end of
2017. Currently, release 15 is in progress under evaluation and
study to give further enhancement on mission critical (MC)-
video [20].
B. LTE-based MCC Systems
The effective performance of LTE- based MCC system will
depend, among many, on the number of User Equipment (UEs)
connected to the eNodeB. An example of a trunked PMR system used
for MCC services shown in Fig. 1. Each first responder’s device is
represented with UE. Some assumptions are made in order to find the
number of MCC user equipment that can be served by one eNodeB.
We assume that resources are not dependent of where the UEs are in
the coverage area of the cell. All of the UEs are connected to the LTE
eNodeB. The MCC model in this study considers only the air interface
to provide connectivity between the UEs and eNodeB. The modeled
LTE cell operates in a 10 MHz bandwidth as assigned for MCC
services. Also, for 2 x10 MHz spectrum, 50 physical resource
blocks (PRB) allocated by 3GPP, it has been derived that the
uplink peak capacity will be 37 Mbps if there are 12 subcarriers,
7 SC-FDMA per symbol and 2 slots [15], [21].
Fig 1. Trunked PMR systems used for MCC services [22].
IV. DEMONSTRATION AND SIMULATION RESULTS
In this study, ARENA software [23] is used to model and
simulate different MCC traffics as shown in Fig. 2. Table I
shows sample of ARENA modules. PMR traffic is being
simulated in ARENA following first in first out (FIFO)
discipline using queue theory [15]. In the create module of
ARENA, the PMR traffic is generated using different
distributions.
Given a traffic arrival rate , (call or message/time), then
the average inter-arrival time (average time between traffic),
is
1
= (1)
For instance, for motion video traffic of 320 x 240 resolution,
10 frames/s, 12 Mbps,
= 12 6 / sec 12 6 / 3600 hrs

Therefore, the average inter-arrival time is:

Fig. 3. Traffic generation setting in the ARENA create module.
1 3600
= = = 0.0003 ℎ . (2)
12 × 10

The average inter-arrival time is used to set the traffic

generation in the create module of ARENA using different
probability distributions and simulations times as shown in Fig.
3.

(a)

Fig. 2. ARENA simulation of MCC traffic model.

(b)
TABLE I ARENA MODULE BLOCK SYMBOLS AND FUNCTIONS
Fig. 4. (a) Traffic processing setting in the ARENA process module (b)
Create module Allow setting resource setting in the resource module of ARENA.
is where traffic processes
generation is set

To assign To mark exit

simulation for incoming
attribute (not traffic
used here)

Fig. 5. Processed traffic exit setting in ARENA.

 Similarly, in the process module of the ARENA as shown in
Fig. 4 a, seize-delay-release attribute is used in order for traffic
to take hold of channel resources for processing or transferring
of traffic and allow queue (delay) for other traffics. The seized
resources are released when processing is done. The service
probability distribution required for processing traffic is also
set with different times required in the process module (e.g
EXP (20) in Fig. 4a). The capacity of resources is set to 50
according to 3GPP standard for MCC (Fig. 4 b). The dispose
module (Fig. 5) marked the exit for the processed traffics or
users.
various traffics for LTE-based MCC uplink network.
From the outcomes of the simulations, the throughput,
is calculated as:
! " #
= × % ! " (& ' ) (3)
! " $ ! %
The number ( ) of traffics that exit the model, number of
traffics generated, resource utilization, and waiting or delay
time given different offered traffics (Mbps), service times, and
simulation times are obtained from the output of the ARENA
software as shown in Figs. 7 and 8. Tables II shows the
simulation results when considering different probability
distributions for traffic generation and service time. Tables III,
IV, and V show the outcomes of the simulations.

Fig 7. Outcomes of simulation from ARENA showing the resource utilization

Fig. 6. Traffic processing setting in ARENA

Fig 8. Outcomes of simulation from ARENA showing traffic in, out and
Furthermore, for running the simulation, the simulation time waiting time
or replication length—represent the time that the LTE-MCC
network can be up and transfer corresponding data
(throughput), is configured in the set up window as shown in
Fig. 6. This study evaluates the cell throughput, user waiting
time, and resource utilization for various MCC users with

TABLE II. RESULTS FOR 5 USERS, 7.2 Mbps EACH, 20 SEC SERVICE TIME, FOR DIFFERENT TRAFFIC GENERATION AND SERVICE
DISTRIBUTIONS

Cases Traffic generation Simulation Traffic Traffic exit Resource Throughput Waiting
distribution / Service time (s) generated utilization (Mbps) time (s)
distribution
1 Lognormal / Lognormal 734 2028 1887 0.98 33.49 16.71
2 Lognormal / Exponential 326 926 983 0.95 30.44 10.73
3 Lognormal / Normal 359 975 832 0.96 30.72 16.38
4 Exponential / Lognormal 410 1166 1022 0.97 31.55 8.05
5 Exponential / Exponential 171 504 361 0.92 25.79 11.34
6 Exponential / Normal 238 670 528 0.95 28.37 11.15
 7 Normal / Lognormal 432 1165 1021 0.94 31.55 5.03
8 Normal / Exponential 1775 4569 4425 0.99 34.87 17.31
9 Normal / Normal 292 784 643 0.95 29.53 12.97

TABLE III. RESULTS FOR 5 USERS, 7.2 Mbps EACH, DIFFERENT SIMULATION TIMES, FOR NORMAL TRAFFIC GENERATION AND
EXPONENTIAL SERVICE DISTRIBUTION

Category Service Simulation Traffic Traffic exit Resource Throughput Waiting

time (s) time (s) generated utilization (Mbps) time (s)
1 10 >> 2000 5124 5104 0.51 35.86 0.00
2 20 1775 4569 4425 0.99 34.87 17.31
3 30 131 339 195 0.91 20.70 13.41
4 40 92 236 92 0.88 14.03 5.45
5 50 80 201 59 0.86 10.57 4.02
6 60 77 196 52 0.86 9.55 3.91
7 120 65 159 19 0.85 4.30 0.12
8 360 61 145 6 0.85 1.49 0.04

TABLE IV. RESULTS FOR 3 USERS, 12 Mbps EACH, DIFFERENT SIMULATION TIMES, FOR NORMAL TRAFFIC GENERATION AND
EXPONENTIAL SERVICE DISTRIBUTION

Category Service Simulation Traffic Traffic exit Resource Throughput Waiting

time (s) time (s) generated utilization (Mbps) time (s)
1 10 >> 2000 5122 5103 0.51 35.87 0.00
2 20 1066 2782 2637 0.99 34.12 16.69
3 30 121 314 171 0.89 19.61 9.83
4 40 93 236 91 0.87 13.88 3.97
5 50 81 206 62 0.86 10.83 3.89
6 60 79 195 50 0.86 9.23 2.98
7 120 65 165 20 0.84 4.36 1.31
8 360 62 152 7 0.84 1.65 2.63

TABLE V. RESULTS FOR 1 USER, 36 Mbps DIFFERENT SIMULATION TIMES, FOR NORMAL TRAFFIC GENERATION AND EXPONENTIAL
SERVICE DISTRIBUTION

Category Service Simulation Traffic Traffic exit Resource Throughput Waiting

time (s) time (s) generated utilization (Mbps) time (s)
1 10 >> 2000 5122 5102 0.51 35.86 0.00
2 20 1249 3217 3071 0.97 34.37 13.43
3 30 131 340 194 0.90 20.54 12.78
4 40 98 242 96 0.87 14.28 7.40
5 50 90 221 74 0.86 12.05 6.38
6 60 81 198 51 0.85 9.27 3.40
7 120 66 166 19 0.83 4.12 0.06
8 360 62 151 6 0.83 1.43 0.00

TABLE VI. MAXIMUM PARAMETER COMPARISON

Approach 3 GPP proposed standard Theory / Reality Simulation

Parameter
Capacity (Mbps) 50 37 36
PRBs 50 48 48
Number of user with = 7.2 & ' 7 5 5
Waiting time (s) 22.89 18.17 17.31
Delay probability 7.00 % 5.55 % 5.29 %

From Tables III, IV, and V, and Figs. 9 and 10, it can be seen
as the simulation time (more time to upload traffic) increases,
the more the traffic throughput transferred. Also, using Normal
traffic generation distribution and exponential service
distribution with a tolerable increase in waiting time gives
more throughput than other probability distribution as seen in
Fig. 11 and Table II. According to Fig. 12, resource utilization
is moderate for when there is one user and it is very high for
when there are three or five users even with same total offered
traffic (36 Mbps). From Figs. 13 and 14 and Tables III, IV, and
V, the throughput is the best when the service time is 20 sec
and simulation time is around 1066 sec and using Normal
traffic generation distribution and exponential service
1
distribution. Also, the waiting time is tolerable for MCC
services. 0.9

0.8
40
1 User, total traffic 36 Mbps 0.7
35 3 Users, total traffic 36 Mbps

Resource utilization
5 Users, total traffic 36 Mbps 0.6

30
0.5
Throughput (Mbps)

25 0.4

0.3
20
0.2
15 1 User, total traffic 36 Mbps
0.1 3 Users, total traffic 36 Mbps
5 Users, total traffic 36 Mbps
10
0
1 2 3 4 5 6 7 8
5 Category

Fig. 12. Resource utilization vs simulation time categories.

0
1 2 3 4 5 6 7 8
Category
18
Fig. 9. Throughput vs simulation time categories. 1 User, total traffic 36 Mbps
16 3 Users, total traffic 36 Mbps
5 Users, total traffic 36 Mbps
40 14

35 12
waiting time (s)
10
30

8
throughput (Mbps)

25
6
20
4
15
2

10
0
1 2 3 4 5 6 7 8
1 User, total traffic 36 Mbps Category
5
3 Users, total traffic 36 Mbps
5 Users, total traffic 36 Mbps Fig. 13. Waiting time vs simulation time categories.
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Simualtion time (s)
2000
Fig. 10. Simulation time vs Throughput. 1 User, total traffic 36 Mbps
1800 3 Users, total traffic 36 Mbps
5 Users, total traffic 36 Mbps
1600
40 20
Throughput (Mbps) 1400
Respective waiting time (s)
Simulation time (s)

1200

30 15 1000

800

600
20 10
400

200

10 5 0
0 50 100 150 200 250 300 350 400
Service time (s)

0 0 Fig. 14. Service time vs simulation time.

1 2 3 4 5 6 7 8 9
Cases
The results of the simulation are also compared with actual /
Fig. 11. Throughput and waiting time vs different cases of probability reality model and 3GPP proposed theory in Table VI and the
distributions.
results of the comparison are shown in Fig. 15. From Fig. 15,
the result of the simulation approach is similar to the reality or
theoretical calculations. Simulating ten users with a traffic
equal to 3.6 Mbps each and a total traffic from all users, which
is about 32Mbps will be uploaded successfully under 6 mins
without bringing down the network and with overall system
delay of 5.96 sec and 0.6 sec for individual delay. Based on the
analysis discussed in this study, it is therefore recommended
that one user should only be allowed to upload at a time if the
user has traffic around 30 Mbps, and ten users can upload to
the eNodeB at the same time if each has a traffic less than 3.6
Mbps. Also, three users with traffic less than 12 Mbps each and
5 users with traffic less than 7.2 Mbps each can simultaneously
use LTE-Based MCC uplink network to connect to the
dispatch center.
45
Proposed approach
40 Theory/Reality
3GPP
35
30
Throughput (Mbps)
25
20
15
10
5
0
1 2 3 4 5 6
Number of users
Fig. 15. Throughput vs number of users.
V. CONCLUSION
This study reports performance of LTE-based MCC using
heterogeneous traffics in simulation scenarios. It is
demonstrated in this study that using Normal traffic generation
distribution and exponential service distribution, delay will be
minimized. Also, the throughput and channel resource are
optimized. The study shows that the lower the traffic volume
per user, the more number of users can simultaneously use the
LTE-based MCC uplink without causing congestion or bring
down the network. This LTE-based MCC network
dimensioning approach can be used in actual implementation
and planning.
Future study will include repeating experiment by varying
simulation time, service time, probability distribution and
number of resources, and evaluate the performance using other
simulation software.
VI. References
[1] ETSI, "TETRA Voice + Data, ETSI, Sophia Antipolis, EN 300392,"
2002.
[2] TETRAPOL Specifications Part - PAS 0001-2, Version 3.0.0,
TETRAPOL Forum, Boisd'Arcy, France," Nov. 1999.
[3] APCO , "Project 25 Standards for Public Safety Digital Radio, APCO
International, South Daytona, FL," Oct. 1995.
View publication stats
[4] P3 Communications, "Evolution of TETRA-"To a 4G All- IP
Broadband Mission Critical Voice Plus Data Professional Mobile Radio
Technology"," November 2011.
[5] D. Jackson, "Public Safety Groups Support LTE at 700 MHz," Urgent
Communications, June 2009.
[6] APCO & NENA , "Endorse LTE as Technology standard for the
Development of Nationwide Broadband Netwrok," APCO News, June
2009.
[7] NPSTC, "National Public Safety Telecommunications Council Press
Release: NPSTC Votes to Endorse LTE Technology for Broadband
Network," June 2009.
[8] USA Government, "Middle Class Tax Relief and Job Creation Act of
2012," Public Law 112-96, 2012.
[9] N. Cackov, J. Song, B. Vujicic, S. Vujicic and L. Trajkovic, "Simulation
and Performance Evaluation o a Public Safety Wireless Network: Case
Study," SIMULATION, vol. 81, no. 8, pp. 571 - 585, 2005.
[10] N. Cackov, B. Vujicic, S. Vujicic and L. Trajkovic, "Using Network
Activity Data to model the utilization of a Trunked Radio System," in
Proc. SPECTS, San Jose, CA, 2004.
[11] H. H. Hoang, R. Malhame and G. Chen, "Traffic Engineering of
Trunked Land Mobile Radio Dispatched Systems," in IEEE Vehicular
Technology Conference, 1991. Gateway to the Future Technology in
Motion, St. Louis, MO, 1991.
[12] H. H. Hoang, R. Malhame and C. Rosenberg, " Communication Load
and Delay in Multichannel Land Mobile Systems for Dispatch Traffic:
a Queuing model Analysis," in Vehicular Technology Conference,
1992, IEEE 42nd, Denver, CO, 1992.
[13] M. Lin, W. Shouming and Q. Wei, "A Novel Traffic Analysis Method
for PoC over LTE Based on Retrial Calling Model," in IEEE 2011
Communications and Networking in China (CHINACOM), 2011 6th
International ICST Conference, Harbin, 2011.
7 8 9 10
[14] A. Ali, M. Alshamrani, A. Kuwadekar and K. Al-Begain, "Evaluating
SIP Signaling Performance for VoIP over LTE Based Mission-Critical
Communication Systems," in IEEE NGMAST, Cambridge, 2015.
[15] K. O. Olasupo, I. Kostanic and T. O. Olasupo, "Analytical Modeling of
LTE-based Network Capacity for Public Safety Communications,"
Universal Journal of Communications and Network, vol. 4, no. 3, pp.
37-46, 2016.
[16] 3GPP, "The Mobile Broadband Standard, LTE," 3GPP, December
2008. [Online]. Available:
http://www.3gpp.org/technologies/keywords-acronyms/98-lte.
[Accessed 22 July 2017].
[17] M. Sauter, From GSM to LTE: An Introduction to Mobile Networks and
Mobile Broadband, John Wiley and Sons,Ltd., 2011.
[18] G. M. Yannick Lair, "3GPP," 20 June 2017. [Online]. Available:
http://www.3gpp.org/news-events/3gpp-news/1875-mc_services.
[Accessed 22 July 2017].
[19] 3GPP TS 22.282 V14.3.0 , "3rd Generation Partnership Project;
Technical Specification Group Services Aspects, Mission Critical Data
Services over LTE (Release 14)," March 2017.
[20] 3GPP TS 22.281 V15.0.0 , "3rd Generation Partnership Project;
Technical Specification Group Services and System Aspects, Mission
Critical Video Services over LTE (Release 15)," June 2017.
[21] M. A. E.-S. M. S. Ayman ElNashar, Design, Depolyment and
Performance of 4G- LTE Networks, John Wiley & Sons, Ltd, 2014.
[22] Anritsu Corp., "Effectively Testing 700 MHz Public Safety LTE
Broadband and P 25 Narrowband Networks," Jan 2013.
[23] Rockwell Automation, "Arena Simulation Software," [Online].
Available: https://www.arenasimulation.com/academic/students.
[Accessed July 2017].