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

net/publication/337488523

ANALYSIS OF GSM NETWORK QUALITY OF SERVICE USING CALL SETUP

FAILURE RATE AND HANDOVER FAILURE RATE INDICES

Article  in  Telecommunications and Radio Engineering · January 2019

DOI: 10.1615/TelecomRadEng.v78.i16.50

CITATIONS READS

2 469

3 authors, including:

Abdoulie Momodou Sunkary Tekanyi Hassan Abubakar Abdulkareem

Ahmadu Bello University Ahmadu Bello University
73 PUBLICATIONS   107 CITATIONS    36 PUBLICATIONS   12 CITATIONS   

SEE PROFILE SEE PROFILE

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

5G NOMA View project

English language studies View project

All content following this page was uploaded by Hassan Abubakar Abdulkareem on 10 March 2021.

The user has requested enhancement of the downloaded file.

 Telecommunications and Radio Engineering, 78(16):1471-1481 (2019)

ANALYSIS OF GSM NETWORK QUALITY

OF SERVICE USING CALL SETUP FAILURE
RATE AND HANDOVER FAILURE RATE
INDICES

A.M.S. Tekanyi, H.A. Abdulkareem, & Z.Z. Muhammad

Ahmadu Bello University, Zaria, 810211, Nigeria

Address all correspondence to: A.M.S. Tekanyi, E-mail: [email protected]

Over the decades, the Global System for Mobile (GSM) communications technology has become
one of the fastest growing and most challenging telecommunications technologies. The Call
Setup Failure Rate (CSFR) and Hand-Over Failure Rate (HOFR) were the most important
Key Performance Indicators (KPIs) used in ascertaining the efficiency of GSM network in
terms of the quality of services rendered. For customers, it is expected that maximum
satisfaction is derived from any service rendered, which was not the case over the years. The
increase of HOFR and poor network availabilities due to increased CSFR became a great
concern to all parties (providers, users, and researchers). Hence, this paper assessed the Quality
of Service (QoS) of MTN GSM network in four geographical areas (Kaduna South, Kaduna
North, Zaria, and Kafanchan) using two Key Performance Indicators (KPIs) of HOFR and
CSFR. The data used in the assessment was collated from MTN Network Management Center
(NMC) with the aid of the FACTS tool. These KPI results were evaluated against those
specified by the Nigerian Communications Commission (NCC) in order to make some
important recommendations from the findings (contributions) to improve the QoS of MTN
network.

KEY WORDS: GSM technology, subscribers, QoS, call setup failure rate, handover
failure rate

1. INTRODUCTION

Over the years, the mobile phone business in Nigeria witnessed tremendous
improvement in terms of better coverage and accessibility, but not with the QoS that
was also expected to be good. The poor QoS was mostly due to increased HOFR as
well as poor network availabilities within the localities and even while moving from
one location to another [1]. The GSM (cellular) network comprises of a Mobile Station
(MS) that connects to the Base Transceiver Station (BTS) via air interface (Abis).
Network performance and QoS estimation are the two most vital aspects serious
considered by mobile operators as the revenue base and client satisfaction are both
directly related to them [2]. GSM communications is the most popular multi-services

0040-2508/19/$35.00 ©2019 by Begell House, Inc. 1471

1472 Tekanyi, Abdulkareem, & Muhammad

base network all around the world. One of these services included data
communications by packet data transport via General Packet Radio Service (GPRS)
and improved data rates for GSM using Enhanced Data for Global Evolution (EDGE).
In additional improvements were achieved when the 3GPP developed the 3rd
Generation (3G) Universal Mobile Telecommunications System (UMTS) standard
followed by 4th Generation (4G), where 4G phones are connected to the Internet for
faster and wider coverage [3]. Network running is not an easy task despite the
equipment that is part of the network. The difficulties of network management are very
much manifested when devices and equipment from different manufacturers are used.
Detecting problems and issues becomes more difficult when circuits in the network do
not have a single console under which they operate [4]. A Network Management
System (NMS) is installed on the network with an online database that is responsible
for the monitoring of all that happens on its network in raw data form. In order
to ensure good network performance with better QoS for customers satisfaction and
retention, this data is analyzed and evaluated to spot events, trends, problem areas, and
drift from recommended KPIs values [1] required for the evaluation of the QoS of a
GSM network. There are several KPIs used in network QoS evaluation, some of which
are network accessibility, service retainability, connection quality, and
network coverage [5]. There are three key techniques commonly used to monitor
network performance, which are drive test, network statistics, and customer
complaints. Client complaints are the most commonly used technique because the
client is always ready to give an input whether valid or not. Though performance
estimation by this technique is easily achievable, it is not the best alternative because
the clients’ experiences can be touching and subjective [6]. A call centre is often
operated through a widespread open workspace for call set/headset connected to a
telecommunication switch with one or more monitoring stations. It can be separately
operated or networked with extra centers, often connected to a business computer
network, including mainframes, microcomputers, and LANs [7]. The drive test
technique used by NCC in 2005 to determine QoS of the GSM networks in Nigeria, is
another technique for performance estimation. Apart from network performance
assessment, it can also be used for the detection of network problem areas, validation
of effects of optimization changes and analysis of root cause of problems in an
operational network [6].

2. MATERIALS AND METHODS

Some of the appropriate materials available for use in this paper were:
i. I Manager M2000 File Transfer Protocol (FTP) that was used to facility data
collection from the network.
ii. Microsoft (MS) Excel tool box was used to plot the data values for easy
understanding and interpretation of infestation results.
To realize the goal of this paper, a step by step procedure was adopted on the data
collected from MTN Network Management Center in line with the established

Telecommunications and Radio Engineering

Analysis of GSM Network Quality 1473

methodology, where data was first analyzed based on KPIs values gathered. The
procedure was as follows:
i. Analysis of the data set.
ii. Determination of Call Setup Failure Rate (CSFR) and Handover Failure Rate
(HOFR).
iii. Evaluation of the performance values.

2.1 Analysis of Collected Data

Analysis was based on the daily evaluation of Base Station Controllers (BSCs), which
were Abuja BSC number 7 (ABHBSC7), Abuja BSC number 13 (ABHBSC13),
Kaduna BSC number 8 (KDHBSC8), Kaduna BSC number 15 (KDHBSC15), Kaduna
BSC number 4 (KDHBSC4), Kaduna BSC number 1 (KDHBSC1), Kaduna BSC
number 3 (KDHBSC3), and Kaduna BSC number 9 (KDHBSC9). Analysis of the data
set was then carried out to investigate the QoS performance for each of the studied
geographical locations. At the Network Management Switching (NMS), a FTP tool
known as the I Manager M2000 was used to draw out the data from the network.

2.2 Determination of Average CSFR

From the two indices of CSFR and HOFR, this study was conducted using the average
variable that was already evaluated using the raw data set obtained from the NMC. The
total average of CSFR and HOFR on a daily basis for a month was obtained by using
MS-Excel tool box. There could be lots of reasons for a poor CSSR [2], amongst them
are the following:
i. Low Signal Strength.
ii. Standalone Dedicated Control Channel (SDCCH) Congestion.
iii. CM Service Reject.
iv. Traffic Channel (TCH) Failure Assignment.
v. Hardware Problem.
The problem of QoS is directly affected by the mobility of subscribers and
difficulty of the radio wave propagation [8,9].
However, most of the network problems are caused by increasing subscribers that
may result in congestion in the network and the changing environment that which may
have harsh conditions impossible to permit radio wave propagation [8,10]. RF
optimization is a continuous process that is required in meeting the best conditions of
service in an evolving network [2].
CSFR was obtained from the formula of Eq. (1) and the average values were
sorted daily for each of the months (January, February, and March, 2016). The CSFR
refers to the number of call setup failure divided by the total number of call attempts. It
is also called the blocking probability and it is expressed in percentage [11,12] as:

CSFR = µ / β × 100%, (1)

Volume 78, Issue 16, 2019

1474 Tekanyi, Abdulkareem, & Muhammad

where µ = Number of Blocked Calls, β = Number of Call Attempt.

2.3 Determination of Average HOFR

Handover failure rate is an important indicator for monitoring mobility. Handover

failure means that mobile station tries to make a handover (inter/intra cell) but for
some reasons it fails [11,12]. These values are kept within some precise threshold in
order to meet the required QoS criteria accepted by the regulatory bodies like Nigerian
Communications Commission (NCC) for subscribers’ satisfaction [8].
It is expressed in percentages as:

HOFR = α / δ × 100%, (2)

where α = Number of Handover Failures, δ = Number of Handover Attempt.

2.4 Evaluation of Performance Values

Performance analysis for the evaluation of the QoS results obtained from the study of
this paper was conducted using Call Setup Failure Rate (CSFR) and Handover Failure
Rate (HOFR) as indices. The performance index values were plotted on graphs for
Kaduna South, Kaduna North, Kafanchan, and Zaria for the month of January,
February, and March, 2016 and measured against the NCC benchmark.
The NCC is empowered to establish minimum QoS standards in service delivery
for the telecommunications industry. These QoS standards ensure that consumers
continue to have access to high quality telecommunications services because basic
minimum quality levels are set for all operators to satisfy in order to remain in
operation (NCC technical standard, 2016). Table 1 gives the NCC KPI benchmarks.

TABLE 1: NCC KPI Benchmarks (NCC, 2016)

Quality Parameter Target Value
Call Setup Failure Rate (CSFR) ≤ 2(%)
Handover Failure Rate (HOFR) ≤ 2(%)

From Table 1, the QoS was measured based on the NCC standard values for each
of the key performance indicators.

3. RESULTS ANALYSIS AND DISCUSSION

Analysis of QoS results using CSFR and HOFR as performance indices is done in this
Section. Graphical representations were used for the plots of daily data values for

Telecommunications and Radio Engineering

Analysis of GSM Network Quality 1475

Kaduna South, Kaduna North, Kafanchan, and Zaria for the months of January,
February, and March, 2016. This is the ability of a subscriber to initiate a call and
granted access. Technically, during a GSM call setup, a speech call is assigned from a
SDCCH to a TCH. If the TCH selected suffers from interference, then the mission
fails. In addition, the mission’s failure message is sent to the Mobile Switching Center
(MSC). The call is then re-established back [13].

3.1 CSFR Results for January, 2016

Figure 1 shows the analysis of January call failure rate. Zaria was completely out of
the threshold restricted limit of NCC throughout the month of January by its inability
to establish calls successfully, except on 14th day of the month. This was an indication
of poor QoS provided to subscribers during the month January, 2016.

FIG. 1: January, 2016 CSFR results for Kaduna North, Kaduna South, Kafanchan, and Zaria

From data analysis most often cells availability reached 100%, but the (PTTCH)
performance was less which indicated that there was high congestion of SDCCH and
the TCH. This made it difficult for any subscriber to have a successful ratio of calls
unblocked to the calls attempted. Kaduna North had the best performance with average
failure rate of 1.43% compared to the NCC benchmark of 2%. Kafanchan also had a
good performance, except on the 6th, 7th, 13th, 14th, and 15th days of the month, which
could have been due to power failure (generator fault, hybrid dc power failure, and
hardware unit’s faults, etc). As for Kaduna South, the failure rate was higher with an
average record value of 2.06% when compared to the threshold level of ≤ 2% set by
NCC. There was frequent fluctuation in the network due to lack of power supply to
service some of the cell sites or Base Transceiver Stations (BTS) in their respective
locations. These cell sites had power failure and a sector was down completely on the
13th, likewise sector E of another cell site registered zero performance for long hours.
All of these resulted into low signal strength for effective coverage of the neighboring
base station, thereby causing a rise in the failure rate.

Volume 78, Issue 16, 2019

1476 Tekanyi, Abdulkareem, & Muhammad

3.2 CSFR Results for February, 2016

Figure 2 shows the CSFR of the four locations during the month of February based on
the statistics obtained from Network Management Centre (NMC) and the graphical
representations. Zaria had the worst performance, with monthly average value of
2.41% that was more than the NCC threshold of ≤ 2%. From the NMC, it showed that
at the beginning of the month 1st, 2nd, 3rd, 4th, 5th, 6th, 7th, 8th, 10th and 11th, except on 9th,
subscribers were unable to gain call access initiated to the BTS through to the MSC as
a result of poor coverage and due to power outages recorded at the base stations and
hardware faults on some of the cells. Kaduna South and Kafancha recorded average
performances were 1.93% and 1.81%, respectively which were within the NCC
benchmark of ≤ 2%. Between 8th to 18th of February and 18th through 30th, there were
severe failure rates in the three locations of Kaduna South, Zaria, and Kafanchan. The
causes of this rise could have been due to TCH congestion, interference, and poor
coverage. On average, Kaduna North had the excellent best service performance of
1.45% CSFR.

FIG. 2: February, 2016 CSFR results for Kaduna North, Kaduna South, Kafanchan, and Zaria

3.3 CSFR Results for March, 2016

Kafanchan area of Fig. 3 had a high rise in failure rate from the beginning of the month
1st to 10th, except for the 3rd that had a sharp improvement, but was unable to sustain
that this availability from 19th to 31st of the month. Network performance was seriously
degraded with the monthly average value of CSFR as high as 2.75%, greater than the
recommended NCC threshold of ≤ 2%. This was as a result of TCH congestion,
frequent power failure, and faulty hardware.
Also in Zaria, the subscribers experienced bad network performance with the
persistent failure rate throughout the month, except for 1st, 3rd, and 4th whose
performances fall within the NCC threshold benchmark of ≤ 2%. From the analysis of
the collated data, hardware unit faults and protracted power failures contributed to the
low PTTCH or cell availability for most of the BTSs which led to poor coverage,

Telecommunications and Radio Engineering

Analysis of GSM Network Quality 1477

thereby impacting negatively on the network. Zaria had an average CSFR value of
2.59% that was more that recommended NCC benchmark of ≤ 2%.

FIG. 3: March, 2016 CSFR results for Kaduna North, Kaduna South, Kafanchan, and Zaria

Considering the recommended CSFR value of ≤ 2% in the results for Kaduna

North, Kaduna South, Kafanchan, and Zaria, Kaduna South had an overall monthly
average performance of 2.07%. Kaduna North maintained its best performance with a
low failure rate of 1.67%. Although, some days like 2nd, 3rd, 17th, 29th, and 31st had
failure rates more than the standard threshold value of NCC. The causes for the rise on
these days were because of TCH congestion, interference, poor coverage due to power
failures at the BTS sites and Faulty hardware units.

3.4 HOFR Results for January, 2016

Handover failure rate as a key performance indicator is the percentage of time that
calls failed to be handed over to the next neighboring BTS or BSC successfully. The
NCC recommended benchmark Handover Failure Rate (HOFR) is set at 2%. This
implies that HOFR value greater 2% indicates poor network quality and poor QoS.
Figure 4 shows the analysis of the HOFR for the four geographical areas studied in the
month of January, 2016 on the assessment of their QoS performances. It was observed
that Zaria had the worst performance on its HOFR, follow by Kaduna North and
Kafanchan, while Kaduna South had the best HOFR record.
From the graph, it is observed that the rise of Kaduna South’s HOFR values above
the 2% benchmark of NCC were recorded on the 2nd, 3rd, 4th, 5th, 20th through 24th, and
28th, to 30th of January. This rise of HOFR values are caused by traffic congestion, link
connection failure, bad radio coverage, incorrect handover ratio, incorrect locating
parameter setting, and high interference. Looking at the Zaria graph, there is sharp rise
above the 2% recommended value of NCC throughout the month of January due to
same factors which affected HOFR of Kaduna South. Kafanchan and Kaduna North
had few and mild sharp rise from the beginning to the end of the month. The causes for
this rise were as a result of TCH congestion, interference, and poor coverage. These

Volume 78, Issue 16, 2019

1478 Tekanyi, Abdulkareem, & Muhammad

four locations studied showed an average HOFR performance for the month of
January, 2016 as follows: Kaduna South recorded 2.07%, Kafanchan had 2.37%,
Kaduna North registered 2.39% and Zaria with worst performance of 3.721%, which
were all greater than the NCC recommended ≤ 2%.

FIG. 4: January, 2016 HOFR results for Kaduna North, Kaduna South, Kafanchan, and Zaria

3.5 HOFR Results for February, 2016

The assessment in Fig. 5 shows that Kafanchan had the best service performance on
the HOFR parameter during February, 2016 when compared to Kaduna South, Kaduna
North, and Zaria. Kafanchan’s performance was better than the rest of the locations,
while Kaduna South was second best on the network’s performance on HOFR.

FIG. 5: February, 2016 HOFR results for Kaduna North, Kaduna South, Kafanchan, and Zaria

Zaria showed bad performance, where there were heavy fluctuation and sharp rises
from the beginning right through 29th February, 2016. The performance was not in line
with the benchmark standard of the NCC. The causes of these fluctuations were due to
traffic congestion, link connection failure, bad radio coverage, incorrect handover
ratio, incorrect location parameter setting, high interference, hardware failure, and the

Telecommunications and Radio Engineering

Analysis of GSM Network Quality 1479

mobile station wrongly measuring signal strength of another co-or-adjacent cell. As

can be seen from the graph Kaduna South had few rises on the 1st, 2nd, 3rd, 4th, 5th, 6th,
and 7th while other days were 14th and 19th of February, 2016. The average monthly
HOFR performances for other locations were as follows: Kafanchan had the average
performance of 1.65 %, Kaduna South had 1.907140956%, that of Kaduna North was
high at 2.54 %, and worst was Zaria with an average value of 3.36 %. The average
performances of Zaria and Kaduna North were not in accordance with the NCC
standard value of ≤ 2% for good QoS, which was an indication of poor network service
in the two areas in February, 2016.

3.6 HOFR Results for March, 2016

Figure 6 shows the grahical assessment representations of HOFRs for Kaduna North,
Kaduna South, Kafanchan, and Zaria for the month of March, 2016. There were great
deviations of handover rate values for all locations studied which were above the
standard ≤ 2% recommended threshold value by NCC.

FIG. 6: March, 2016 HOFR results for Kaduna North, Kaduna South, Kafanchan, and Zaria

Only Kafanchan had a better QoS on handover rate from 10th to 22nd of March and
between 25th-31st March,2016 with a total monthly average handover rate performance
of 1.92%, which was within the NCC benchmark. Looking at Kaduna North from the
graphical analysis, there were high handover failure rates throughout the month, with
an average performance HOFR value of 2.71%, way above the threshold limit of ≤ 2%
stipulated by NCC.
Also, Zaria’s performace was very poor, with high rise in HOFR value of 3.09%
well above the NCC benchmark recommended value of ≤ 2% which implies that MTN
subcribers had a bitter experience on QoS in this Zaria location . Similarly, it was
discovered that the causes for the rise in HOFR values and their fluctuations were due
to traffic congestion, link connection failure, bad radio coverage, incorrect handover
ratio, incorrect location parameter setting, and high interference, co-channel or
adjacent-channel service, which resulted to improper handshaking during call

Volume 78, Issue 16, 2019

1480 Tekanyi, Abdulkareem, & Muhammad

establishment from one location to another. Kafanchan had the best performance,
followed by Kaduna South with 2.23% and worst for Kaduna North and Zaria.

4. RECOMMENDATIONS

In order to improve the network performance and the quality of service, it is

recommended that the management of MTN Nigeria and the government NCC
regulatory body should ensure the following:
1. More base stations are installed and carry-out Radio Frequency (RF)
optimization often on antennae to ensure total coverage and successful handover
of calls in the network.
2. The federal government, NCC, and other regulatory bodies should enforce a
regular monitoring schedule on networks by operators and performance reports
at very short intervals are filed for inspection by NCC so that Quality of Service
(QoS) is established to meet up demands of subscribers.
3. Similar researches should be carried out in other parts of the country using the
same or other QoS parameters to give a more detail picture of the network trend
in the country as a whole.
4. MTN, in particular, should ensure that all traffic affected alarms on the BTS
such as Voltage Standing Wave Ratio (VSWR), Receive Signal Strength
Indicator (RSSI), etc. are cleared to reduce call setup failure rate.

5. CONCLUSION

In general, Call Setup Failure Rate (CSFR) is caused by the following reasons which
also contribute to unpleasant experience subscribers subjected to when NCC
benchmark values are not respected.
i. Low Signal Strength.
ii. Interference.
iii. High Congestion on Standalone Dedicated Control Channel (SDCCH).
iv. High Congestion on Traffic Channel (TCH).
v. Transmission Fault.
vi. Faulty Transceiver Unit (TRU)/Hardware.
vii. Central Processor Overload.
viii. Software File Congestion.
ix. Cell is not defined in Mobile Switching Center (MSC).

REFERENCES

1. Ukhurebor Kingsley Eghonghon, (2017) Evaluation of the quality of service of a cellular network
using the network statistics, International Journal of Advanced Engineering and Technology, ISSN:
2456-7655, 1(5), pp. 01-07, www.newengineeringjournal.com.

Telecommunications and Radio Engineering

 Analysis of GSM Network Quality 1481

2. Venkata-Sai Sireesha, B., Varadarajan, S., Vivek, and Naresh, (2015) Increasing of Call Success
Rate in GSM Service Area using RF Optimization, Engineering Research and Applications (IJERA),
ISSN: 2248-9622, 1(4), pp.1479-1485.
3. Rajesh Kumar Upadhyay, Vijay Kumar Singh, and Rajnish Kumar, (2014) Performance Analysis of
GSM Network, International Journal of Advance Research in Science and Engineering (IJARSE),
3(5), pp. 244, ISSN-2319-8354(E) www.ijarse.com.
4. Eleni Rozaki, (2015) Design and Implementation for Automated Network Troubleshooting using
Data Mining, International Journal of Data Mining & Knowledge Management Process (IJDKP),
5(3), pp 9-27.
5. Lawal, B.Y., Ukhurebor, K.E., Adekoya, M.A., and Aigbe, E.E., (2016) Quality of Service and
Performance Analysis of a GSM Network in Eagle Square, Abuja and Its Environs, Nigeria,
International Journal of Scientific & Engineering Research, ISSN 2229-5518, 7(8), pp. 1992-2003.
6. Ukhurebor, K.E, Awodu, O.M, Abiodun, I.C, and Azi, S.O., (2015) A Comparative Study of the
Quality of Service of GSM Network during Crowd Upsurge in University of Benin Nigeria,
International Journal of Scientific & Engineering Research, 6(10), ISSN 2229-5518.
7. Sheik Manzoor, A.K. and Shahabudeen, F., (2014) A Study on Key Performance Indicators and their
Influence on Customer Satisfaction in Call Centers, International Journal for Engineering Research
(IJER), Serials Publications, 11(2), pp. 1-6, ISSN: 0972-9380.
8. Gordon Ononiwu1, Bukola, O.H., Akinwole2, Cosmos Agubor1, and James Onojo, (2016)
Performance Evaluation of Major Mobile Network Operators in Owerri Metropolis of Nigeria,
International Journal of Emerging Technologies in Computational and Applied Sciences (IJETCAS),
ISSN (Print): 2279-0047, ISSN (Online): 2279-0055, pp. 06-13, www.iasir.net.
9. Shoewu, O. and Edeko, F.O., (2011) Outgoing Call Quality Evaluation of GSM Network services in
Epe, Lagos State, American International Journal of Scientific and Industrial Research, SSN 2153-
649XZ, pp. 409-417.
10. Jatin, Karishan Kumar, (2016) Study and analysis of call dropping and handover problem in cellular
system, International Journal of Advanced Research in Computer Engineering & Technology
(IJARCET), ISSN: 2278-1323, 5(6), pp. 1776-1.
11. Kollar, M., (2008) Evaluation of Real Call Set up Success Rate in GSM, Acta Electrotechnica etc
Informatics, 8(3), pp. 54-56. ISSN 1335-8243 © 2008 FEI TUKE.
12. Idigo, V.E., Azubogu, A.C.O., Ohaneme, C.O., and Akpado, K.A., (2012) Real-Time Assessments of
Quality of service of Mobile Cellular Networks in Nigeria, International Journal of Engineering
Inventions, 1(6), pp. 64-68.
13. Nnochiri Ifeoma, U., (2015) Evaluation of the Quality of Service of Global System for Mobile
Telecommunication (GSM) Operators in Nigeria, Journal of Multidisciplinary Engineering Science
and Technology (JMEST), ISSN: 3159-0040, 2(7).

Volume 78, Issue 16, 2019

View publication stats