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

net/publication/353175137

PERFORMANCE ANALYSIS OF GSM NETWORK

Article · May 2014

CITATIONS READS
18 1,447

3 authors, including:

Rajesh Kumar Upadhyay

Mangalayatan University
10 PUBLICATIONS 23 CITATIONS

SEE PROFILE

All content following this page was uploaded by Rajesh Kumar Upadhyay on 12 July 2021.

The user has requested enhancement of the downloaded file.

International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

PERFORMANCE ANALYSIS OF GSM NETWORK

1
Rajesh Kumar Upadhyay, 2Vijay Kumar Singh, 3Rajnish Kumar,
1, 2
Department of Electronics & Communication Engineering,
Mangalayatan University, Aligarh, UP (India)
3
Reliance Communication, UP West & Uttarakhand Circle, Meerut, UP (India)

ABSTRACT

Global System for Mobile (GSM) is a second generation cellular standard that was developed to solve the
fragmentation problems of the first cellular systems and it specifies digital modulation techniques and network level
architectures and services. To keep up with the change in the technology, organizations must be a learning
organization to keep them in the competitive level. My paper is based on Performance Analysis of GSM Network
(Reliance GSM) in Aligarh City, UP, India. Two approaches are used to evaluate the network performance, namely:
key performance indicators (KPIs) and drive test. In this paper, according to KPI reports and previous drive test
samples, necessary changes are done in network parameters for getting the improved value of TCH resources, hand
over success rate, call setup success rate, Rx level and Rx quality in acceptable ranges for better utilization of
available resources for providing best quality of service for customers.

Keywords: Global System For Mobile; Gsm; Mobile Network, Mobile Network Optimisation; Drive
Test; Key Performance Indicators; Kpis; Tch Drop Rate; Handover Success Rate; Signal Quality;
Signal Level; Reliance.

I INTRODUCTION

In today's world communication technology plays a vital role. It has become an important tool for sharing
information in a personal life or in any organizations for the transfer of data. The telephone was introduced to the
public in 1876. Although the wire bound tele-service was growing at that time, there was a need of wireless
communication for army, police etc. In the late 1940’s, Frequency Modulation (FM) push to talk on the telephone
were used to connect with the public telephone network. This analog system used a single high power transmitter
and a large tower to cover the distance of 50 km. It used the 120 KHz of radio frequency bandwidth in a half duplex
mode. These systems were unable to support the large number of subscribers because of spectrum shortage.
Around the world different analog cellular systems were developed in later years in 34 different countries. These
analog systems soon became inefficient because a mobile telephone developed for one system cannot be used with
another so roaming was limited. In order to overcome these problems, the conference of European Posts and

244 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

Telecommunications (CEPT) formed. In 1982, a Group Special Mobile (GSM) was developed which was using a
digital technology. In 1989, GSM responsibility was transferred to European Telecommunication Standards Institute
(ETSI), and phase 1 of the GSM specifications was published in 1990. Commercial service was started in the mid of
1991 with GSM 900 which uses 25MHz and band width 890-9l5 MHz from mobile to base station and 935-960
MHz from base station to mobile). With the development of digital telephony GSM became popular due to these
features:
 Superior speech quality Low terminal and service costs
 International roaming under one subscriber directory number.
 ISDN compatibility
 High level of security
 Support new range of services and network facilities

Among the existing cellular networks, global system for mobile (GSM) communications is the most popular cellular
communication system all around the world [1, 2]. This system was developed over time to include data
communications by packet data transport via general packet radio services (GPRS) and enhanced data rates for GSM
evolution (EDGE). Further improvements were made when the 3GPP developed third generation (3G) UMTS
standard followed by fourth generation (4G) LTE advanced standard [3]. In recent years, a great deal of attention
has been paid to the planning, evaluation and optimisation of mobile cellular networks [4, 5]. There are several
research papers in the literature that addressed the evaluation and optimisation of operational GSM networks. Thus,
Key performance indicators (KPIs) to evaluate the performance of an operational GSM network. The network
performance evaluation is based on four major KPIs, i.e., call setup success rate, call drop rate, handover success
rate and traffic congestion. Every KPI is explored and improvement methodologies are suggested [6, 7]. However,
these improvement methodologies were not implemented in the network. Evaluation of the handling of traffic and
network utilization of an operative GSM network during eight months may a basic time span for a network. They
developed a regression-based forecasting model for the traffic [8]. The performance of GSM and GPRS operational
network is presented with a review of the most common KPIs that are used to evaluate the performance of GSM and
GPRS networks [9]. The relationships between these KPIs are introduced and thresholds for some KPIs are
suggested so that the GSM operators should not exceed them. Addressed the optimisation of GSM network, data
sorting and analysing, implementing the optimisation and system fine tuning [10]. Network dimensioning such as
BSC, MSC and other related parameters are addressed and the top ten wireless parameters are listed which are the
most important from the authors’ point view. KPIs for quality of service (QoS) evaluation in GSM network are
identified. Four assessment parameters network accessibility, service retainability, connection quality, and network
coverage) for evaluating the QoS were applied on four GSM networks in Nigeria [11]. The results of that study
showed that the QoS of GSM networks in Nigeria is unreliable and the network accessibility and retainability are
unsatisfactory. Aim of study for presenting the QoS of network optimisation and evaluation of KPIs provided by
GSM operators in terms of the ability to establish and maintain call connections, call retention, handover, inter and

245 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

intra network call set-up [12]. Proposal of a measurable indicator of quality of experience (QoE) in EDGE networks,
called busy hour throughput per user, which can be employed to direct practical network design and optimisation
[13]. Semi-automatic tool is developed to perform the network optimisation by tuning the parameters of a live GSM
network. The authors select the most common KPIs to evaluate the QoS and they assign thresholds for these KPIs.
The developed tool is then used to suggest some recommendations that can be applied to the cell so as to enhance its
performance [14]. Evaluation of the call set up success rate as one of the most important KPI in GSM system.
Suggestion for calculating this KPI as the ratio of the assigned traffic channels (TCHs) to the channel requests [10].
From the literature survey above, some authors have suggested several approaches to improve the available KPIs.
However, most of these approaches were not implemented and evaluated for a live GSM network. Other authors
have used the drive tests to evaluate the performance of active GSM networks in terms of QoS. Nevertheless, to the
best of our knowledge, there are no comprehensive research works that combine both the KPIs and the drive tests to
evaluate and optimise the performance of operational GSM mobile networks.
In this paper, a comprehensive research study is carried-out on the performance evaluation and optimisation of
Reliance GSM mobile network in Aligarh City is selected as a case study. Our contribution is twofold: Firstly, both
the KPIs and the drive tests are used to evaluate and optimise the performance of the GSM network in Aligarh City.
In addition, operation and support subsystem (OSS) optimisation tools are used in the final optimisation process.
Secondly, performance of the GSM network should be better as before. Figure 1.1 illustrates the various steps
carried-out in this research. Starting from the operational Reliance GSM network, an evaluation step is performed by
using the data from network audit, KPIs and drive tests. After the evaluation step, an optimisation step is performed
followed by implementation. The optimisation process is cyclic till the best performance for the network is reached.
At that point a comparative study is carried-out between the performance of the GSM network before and after
introducing the new implementation. This research is conducted from October 2013 to April 2014.

Fig. 1 Various Steps Carried-Out in This Research Paper

246 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

II. GSM NETWORK AUDIT IN ALIGARH CITY

2.1 Basic Network Information
The selection of Aligarh City to perform this research for minimizing the TCH drop rate and improving the QoS.
Table 1 shows the basic information about Reliance GSM network in Aligarh City of October 2013. This
information includes active sites and cells, serving base station controller (BSC), serving Mobile Switching Centre
(MSC), broadcast control channel (BCCH) and TCH frequencies and total number of transmitter-receiver units
(TRUs). Figure 2 is a map from Map Info Professional software that illustrates all active and planned sites with cell
directions in Aligarh City.
Table 1 Basic Information of Reliance GSM, Aligarh
Detail Specification
Operator Reliance GSM
Frequency Band GSM 1800 MHz
Total Sites 66 Sites
Total Cells 198 Cells
TRX 594
Total BCCH & TCH 43
Serving BSC RC-G-ALGH-BSC-01-HU
Serving MSC Aligarh MSC

Fig. 2 Reliance GSM sites in Aligarh City, UP displayed on MapInfo Professional

247 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

2.2 Frequency Reuse Plan

Although there are 373 channels in the GSM 1800 band, Reliance GSM network in Aligarh has only been allocated
43 channels. These channels are divided into traffic and control channels. More particularly, the available 43
channels can divided into 17 TCHs and 26 BCCHs. The absolute radio frequency channel number (ARFCN) of the
BCCH channels and TCH channels are allocated according to frequency planning and requirement to accommodate
198 cells.

III. KEY PERFORMANCE INDICATORS

KPIs are used to evaluate the performance of an operational GSM network. These KPIs come in form of counters
from the OSS of GSM network. These counters are then converted into a more readable way. The most common
KPIs are listed below:
1. TCH drop rate: this is one of the most critical KPIs in GSM networks since it is annoying to the customer as
well as the operator. It is the percentage of lost connections to the total number of connections for a given cell in the
active mode (on call). The value of this KPI ranges from 2% in the initial launching for the cell and it must be
decreased to 0.5% in normal operating conditions.
2. Standalone dedicated control channel (SDCCH) drop rate: It is the percentage of the lost SDCCH connections
to the total SDCCH connection attempts. SDCCH is used during vital roles such as call setup and mobile
registration.
3. TCH congestion and congestion perceived by subscriber: There are two different approaches when dealing
with TCH congestion. The TCH congestion perceived by subscriber is more realistic since it is from user’s point of
view rather than the congestion observed by the network and it is given by TCH congestion.
4. TCH assignment success rate: It is a measure of the successful TCH assignments to the total attempts. It is
recommended to keep this value as close to 100% (typical value is 98 %).
5. Handover success rate: It is the percentage of successful handovers to the total number of handover attempts.
6. SDCCH congestion: SDCCH is used during call setup, mobile registration and SMS. The SDCCH congestion is
the percentage of the time that all SDCCH resources are busy within a given cell. The accepted value is no more
than 0.5.
7. Call Setup Success Rate (CSSR): It is a measure of call setup success rate of attempted calls by subscribers in a
cell and it should me minimum 98 % in a good network.

IV DRIVE TEST AND KPI RESULTS

Two approaches to evaluate the performance of a live GSM network, namely KPIs and drive tests. On the one hand,
KPIs give us detailed statistics for many events in GSM cells. On the other hand, drive tests show the realistic
experience of the customer. Customer’s point of view is more important since it reflects the real life scenarios.

248 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

TEMS investigation data collection Version 10.0.5 is used for drive testing of GSM network all over Aligarh City.
This drive test was performed in Oct 2013. TEMS investigation route analysis and MapInfo Professional for
analyzing these collected drive test log files. The advantages of the drive test are: Firstly, drive test is a powerful
tool for the radio frequency (RF) analysis and problem solving. Secondly, scanner tool used in drive test is a very
good tool for detecting interfering signals.
Finally, the drive test gives the exact geographical location for each sample through the connected GPS receiver.
There are many parameters that are gathered by the drive test. The most important parameters are the received signal
level (RXLEV) and the received signal quality (RXQUAL). RXLEV is the received signal power in units of dBm.
The minimum acceptable RXLEV is commonly set to –104 dBm for the cell. While RXLEV value of –60 dBm is
considered as excellent value. For a customer to be able to access the GSM network, it is strongly recommended that
there will be a sufficient coverage with acceptable quality. RXQUAL is a measure of the QoS and it is given by
GSM scale from 0 to 7, where 0 is the best quality and 7 is the worst. TEMS uses RXQUAL to identify the quality
of a voice call or data session during the dedicated mode. RXQUAL is used to estimate the quality of GSM network
service measured in terms of Bit Error Rate (BER) before channel decoding. Table 2 shows how the RXQUAL is
mapped to the BER percentage.
Table 2 RXQUAL with corresponding BER values
RXQUAL BER
0 <0.2%
1 0.2% to 0.4%
2 0.4% to 0.8%
3 0.8% to 1.6%
4 1.6% to 3.2%
5 3.2% to 6.4%
6 6.4% to 12.8%
7 >12.8%

In October 2013, a drive test was carried-out in the streets of Aligarh City. The distributions of the collected signal
samples of RXLEV and RXQUAL are shown in Figure 3 and 4 respectively after optimization of network.
According to Fig 3, most of the samples are within the interval from -85 to 0 dBm with average of –85 dBm. Note
that the number of samples with RXLEV less than –85 dBm are negligible. Thus, the signal level in Aligarh City is
very good. A sharp spike around the value of –85 dBm comes from the fact that the drive test car stuck in traffic jam
close to the RF site in the city centre such that many samples are taken with very good RXLEV. According to Fig. 4,
about 68% of the samples have excellent quality corresponding to RXQUAL value of 0. In addition, about 3% of the
samples have poor quality corresponding to RXQUAL values of 5 and 7. To analyze the data from drive test, four
coverage classes are defined interms of RXLEV and RXQUAL according to Table 3.

249 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

Fig. 3 Fig. 4

Fig. 3 RX Level Full (dBm) Distribution of collected samples from drive test

Fig. 4 RX Quality Sub distribution

Table 3 Coverage levels legend in terms of RX LEV and RX QUAL

Coverage class Condition

Level 1 RX LEV> = –70 and RX QUAL< = 2
(RX LEV> = –85 and RX QUAL< = 4)
Level 2 and
(RX LEV< –70 or RX QUAL>2)
(RX LEV> = –100 and RX QUAL< = 6)
Level 3 and
(RX LEV< –85 or RX QUAL>4)
Level 4 RX LEV< –100 or RX QUAL>6

Thus, level 1 is the best level since it guarantees the best of both RXLEV and RXQUAL. Level 2 provides the users
with good RXLEV and good RXQUAL. Level 3 provides the user with acceptable RXLEV and RXQUAL.
However, level 4 is the worst level since it combines poor RXLEV with poor RXQUAL. Table 4 & 5 shows the
percentage distribution of both parameters as obtained from the data collected during the drive test according to
operator legends. Figures 5 and 6 show, respectively, the drive test maps of RX LEV and RX QUAL in Aligarh
City. This is important since it shows the geographical location for each sample. By inspecting drive tests of each
area in Aligarh City, poor performance is negligible after optimization.

250 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

Table 4 RX level distribution in drive test in Aligarh City

Table 5 RX Quality distribution in drive test in Aligarh City
Table 4 Table 5

RX Level Full Value (dBm) No. of Samples

% of samples
RX Quality Sub
-85 dBm to 0 dBm 29480
0 to 1 68
-95 dBm to -85 dBm 437 2 to 5 29
5 to 7 3
-95 dBm to -110 dBm 1

Fig. 5 Rx Level Full dBm Plot of Aligarh City Fig. 6 Rx Quality Plot of Aligarh City

During optimization process, two stages which are the manual optimisation process and the automatic optimisation
process using the OSS optimisation tools. The manual optimisation process starts with capacity analysis and
neighbour relation plan. It yields two outcomes which are new frequency plan for both BCCH and TCH frequencies,
and new neighbour relation plan. The outputs of the manual optimisation process are then fed to the OSS
optimisation tools. The main output of the second stage is a frequency plan for BCCH and TCH channels with
reduced interference. An enhanced neighbour relation plan with better handover performance is the other outcome
of this stage. We also get good KPI results and some of them are shown in Table 6.

251 | P a g e
www.ijarse.com
International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

Table 6 Drive Test & KPI Reports

Achievement (After
Parameter Target Optimization)
Technology GSM
Band of operation (MHz) 1800
Route covered (km) 400
No. of Sites 66
% of route with good outdoor coverage(Rx level/M Tx) >-85dBm/<0dBm 99%
TCH Drop <2% 0%
CSSR (%) >=98% 100%
Call Drop (%) ≤ 1.5% 0%
Rx Quality DL(<=5) %/FER %(<=2) >=97% 97%
Handover Success Rate (%) >=96% 99.84%

V CONCLUSION
This paper describes a performance analysis and optimisation of Reliance GSM Mobile Network in Aligarh City,
UP, India. Two methods are used to evaluate the performance of the network, namely: drive test and KPIs.
Performance of the network before and after optimisation showed significant performance improvement in terms of
TCH drop rate, handover success rate, received signal level and received signal quality, call setup success rate.
Indeed, the average TCH drop rate is reduced 0 %. In addition, the average handover success rate is increased to
99.94 %. Furthermore, the percentage of samples poor signal strength and poor quality reduced to 1.00 % and 3.00
% respectively that are desirable for proper coverage to the customer’s satisfaction in term of signal strength and
good quality of service.

VI ACKNOWLEDGEMENTS
The authors would like to thank the Reliance Communication, UP West & Uttarakhand Circle (India) for their
support in this research project.

REFERENCES
[1]. Rappaport, T.S. (2001) Wireless Communications: Principles and Practice, Prentice Hall, USA.
[2] Mkheimer, Baha and Jamoos, Ali (2012),’ Evaluation and optimisation of GSM network in Jenien City,
Palestine’ International Journal Mobile Network Design and Innovation, Vol.4, No. 4, pp. 201–213
[3]. Dahlman, E., Parkvall, S. and Skold, J. (2011) 4G: LTE/LTE-Advanced for Mobile Broadband, Academic Press,
UK.
[4]. Mishra, A.R. (2004) Fundamentals of Cellular Network Planning and Optimization 2G/2.5G/3G... Evolution to
4G, Wiley & Sons, Ltd., England.

252 | P a g e
www.ijarse.com
 International Journal of Advance Research In Science And Engineering http://www.ijarse.com
IJARSE, Vol. No.3, Issue No.5, May 2014 ISSN-2319-8354(E)

[5]. Mishra, A.R. (2007) Advanced Cellular Networks Planning and Optimization2G/2.5G/3G & Evolution to 4G,
Wiley & Sons, Ltd., England.
[6]. Haider, B., Zafuaalh, M. and Islam, M.K. (2009) ‘Radio frequency optimization & QoS evaluation in
operational GSM network’, in Proceedings of the World Congress on Engineering and Computer Science, Vol.
1, pp. 393–398, San Francisco, USA.
[7]. Kollar, M. (2008) ‘Evaluation of real call set up success rate in GSM’, Acta Electrotechnical Informatics, Vol.
8, No. 3, pp.53–56.
[8]. Bouamalf, M., Abrache, J., Aouam, T. and Harroud, H. (2009) ‘Traffic analysis for GSM networks’, in
Proceedings of the IEEE International Conference on Computer Systems and Applications, pp.498–503.
[9]. Kyriazakos, S., Gkroustiotis, V., Karambalis, C., Kechagias, C., Papaoulakis, N., Nikitopoulos, D. and Karetsos,
G. (2002) ‘Performance evaluation of GSM and GPRS systems based on measurement campaigns’, in
Proceedings of the 5th IEEE International Symposium on Wireless Personal Multimedia Communications, Vol.
3, pp.1167– 1171.
[10]. Jia, S., Guo, Y., Wang, Z. and Kang, K. (2010) ‘Some lessons learned in GSM network optimization’, in
Proceedings of the IEEE International Conference on Machine Vision and Human-machine Interface (MVHI),
pp. 569– 571.
[11]. Popoola, J.J., Megbowon, I.O. and Adeloye, V.S.A. (2009) ‘Performance evaluation and improvement on
quality of service of global system for mobile communications in Nigeria’, Journal of Information Technology
Impact, Vol. 9, No. 2, pp. 91–106.
[12]. Shoewu, O. and Edeko, F. O. (2011) ‘Outgoing call quality evaluation of GSM network services in Epe,
Lagos State’, American Journal of Scientific and Industrial Research, Vol.2, No. 3, pp. 409–417.
[13]. Gao, F., Gao, Z., Wen, L., Zhang, B., Tang, D. and Hsu, D. (2009) ‘A research on key performance indicator
of measurable QoE of EDGE network’, in Proceedings of the IEEE International Conference on Information
Engineering and Computer Science (ICIECS), pp. 1–4.
[14]. Selmi, A.H. and Tabbane, S. (2004) ‘For a GSM network tuning’, in Proceedings of the IEEE International
Conference on Control, Communications and Signal Processing, pp. 527–530.

253 | P a g e
www.ijarse.com

View publication stats