Comparative Analysis of the Performances of …...Combined model, the perfor mance of the GSM networ k is extensively evaluated using the key performance indicators which include;

Copyright © 2015 SciResPub.

Comparative Analysis of the Performances of aCombined and Non Combined Models Designs ForCongestion Controls in Global System For Mobile

Communication (GSM) Network1MBACHU, C. B., 2USIADE, R E.

1 Department of Electrical/Electronic Engineering, Anambra State University Uli, Nigeria2 Department of Computer Engineering, Delta State Polytechnic Otefe-Oghara, Nigeria

Email: 2 [email protected], [email protected]

ABSTRACT

In this resear ch work, a Co mbined Model fo r managing congestion in G SM network based on CallPriority, Handoff Call Buffer and Frequently Recent Call concept s is de veloped. Based on thi sCombined model, the perfor mance of the GSM networ k is extensively e valuated using the keyperformance indicators which include; Call Set-up Success Rate (CSSR), Call Drop Rate (C DR), CallCompletion Success Rate (CC SR) and Traffic Channel Congestion Rate (TCHCR). After theevaluation, the various values of these parameter indicators namely CSSR, CDR, CCSR, T CHCR areapproximately 98%, 2%, 97% and 2% respectively which satisfies the Nigeria CommunicationCommission (NCC) acceptable bench mark for mobile operators. This paper also discusses differentdeveloped schemes which include Dynamic Channel Allocation (DCA), Automatic Call Gappin g(ACG), Adaptive Call Admission (ACA) and Call A dmission Control (CAC) for c ontrollingcongestion and pro viding good Quality of Service (QoS) in G SM networks. A comparison of theperformance of this Co mbined Model with the existing system based on mobile cellular network KeyParameter Indicators as mentioned above reveals that this model has an improved performance (Qualityof Service) over the existing system. Finally, the contribution of this re search work is to make thisdesigned (Combined) model a more reliable al ternative to existing models for managing congestion inGSM network.

KEYWORDS: Global System for Mobile Communication GSM; Mobile Network Evaluation; DriveTest; Key Performance Indicators KPIs; Quality of Service QoS. Call Prioritization, Handoff Call.

International Journal of Advancements in Research & Technology, Volume 4, Issue 10, October -2015 ISSN 2278-7763 109

IJOART

IJoART


1. Introduction

In recent time, especially the 21st century, information and communication technology has formed the

bedrock of technological and ec onomical growth in the world. The reason is not far-fetche d. People

need to comm unicate, that is, they desire to reach out to others and share information, ideas and

resources. This desire has been a dri ving force, motivating man to continuously seek for a new and

more effective means of disse minating information to one another on real tim e basis irre spective of

distance [1].

The analogue system of teleco mmunication was associated with l ots of setback or li mitations which

made it impossible for subscribers to get the desired satisfaction for its available services. The two

major inherent limitations of the analogue cellular systems are severe confined spectrum allocation and

incompatibility among the various analogue services available [2]. This consequently l ed to the

convergence of the Europeans on a uniform standard for second generation digital system called global

system for mobile teleco mmunication (GSM). The GS M network is a mobile telecommunication

technology system that uses the Time Division Multiple Access (TD MA) to di vide the channel into

time slots. It offers hi gh quality voice communication and low bandwidth (9. 6kb/sec) data connection

for fax, short message service (SMS) and full dial-up conn ection to the internet for e-mail and web

browsing, usually requirin g a mobile computer or intellig ent handset. Increase in de mand for GSM

services may lead to congestion if there is no commensurate increase in congestion control measures.

In digital communication congestion occurs when the nu mber of subscrib ers attempting to

simultaneously access the networ k is more than the capacity the network can handle or sustain.

According to [3], it is the unav ailability of network to the subscriber at the ti me of making a call.

Congestion can also be defined as a sit uation when a subscriber cannot obtain a con nection to the

wanted subscriber i mmediately [4]. In another related publication, [5] describ ed congestion as a

situation that arises when the num ber of calls emanating or terminating from a particular n etwork is

more than the capacity the network is able to cater for at a time.

A lot of research works and publications have been carried out and written respectively in different

congestion control measures. Some of these include [6] dynamic channel allocation (DCA) where there

is no fixed channel. A ll channels are kept in a central pool and are assi gned dynamically to radio cells

as new calls arrive in the system. After a call is terminated, the channel is returned to the common pool.


IJOART

IJoART


Since generally more than one channel m ay be available in the central pool to b e assigned to a caller

that requires a channel, the system fails to develop strategy that must be applied to select th e assigned

channel. The authors in [ 7] did an appraisal of the performance of GSM operators in a country known

as Nigeria in African Continen t. Having evaluated the parameters that attrib uted to poor quality of

service by o perators, they ca me up with methods that are su ggested towards improving network

performance. A hybrid model was developed in [8] for congestion management which was a

combination of other models. However, they failed to make provision for managing handoff calls when

network channels are unavailable. This leads to high blocking probability for handoff/handover calls.

A work in [9] proposes the concept of prioriti zation of handoff calls over new calls by usin g buffering

technique since it is desirable to handle an ongoing call and to accept new ones when the bandwidth

has reached its full capacity utilization. Similarly, another research by [10] proposed a threshold-based

guard channel policy. The polic y allocates some channels to handoff calls when the number of busy

channels exceeds the given threshold thereby blockin g new calls. In a related sche me which was

proffered by [11] it m akes use of bufferin g calls that can tolerate delay if no channels are free. The

major setback associated with the last thr ee works, which are guard channel schemes, is their inability

to establish a record of frequently recent rejected/blocked calls in order to determine the order of their

access to network facilities when the channel is free. In addition, these schem es perform better in light

traffic only.

A scheme where calls are queued and no ne w calls are granted access before the pre vious calls in the

queue is presented in [12]. It is a s tricter scheme than the previously discussed guard channel ones.

This scheme is not effecti ve because the rate of call rejection/blo ck is very hi gh. Another scheme

performed a co mbination of queuin g of new calls and guard channels. The results s howed that the

blocking of handoff calls decreases much faster as the queuing probability of new call s increases. A

scheme that combined ad mission control and bandwidth adaptation to enhance the Quality of Service

(QoS) provision was developed in [13]. Another proposed tech nique for con gestion control was

developed to adopt the principle of Automatic Call Gapping (ACG) which helps to reduce call attempts

by allowing only one call attempt per specific gap interval [14]. This scheme was not found to be very

effective because it per mitted a lot of idle t ime. A congestion control scheme for wireless mobile

network, which had a combination of call admission control with buffer management, was pr esented in


IJOART

IJoART


[15]. The authors in [16] and [17] sugg ested upgrading of existing facilities, installation of additional

base stations and switchin g centres, and where applicable impro ving power supply to control

congestion in GSM networ k. A combined scheme which incorporat ed the Adaptive Call Admission

(ACA) scheme and load balancin g strategy was developed in [18] and the sche me is c apable of

minimising the New Call Bloc king Probability (NCBP) and the Handoff Call Droppin g Probability

(HCDP). In [5] and [4], the researchers attributed the causes of congestion in GSM network in some

countries like Nigeria to factors such as exceeding the carrying capacity of network facilities, use of

mobile phones for data transfer and multi media activities, vandalisation of networ k equipment and

unfavourable weather conditions.

In this pap er, the proposed research de velops a new sche me that can control con gestion in GSM

network. The network performance evaluation is based on four m ajor key parameter indicators (KPIs)

which include call setup success rate, call drop rate, han dover success rate and traffic channel

congestion rate. Every KPI is explored and improvement methodologies are suggested. These include a

combination assignment of priorit y levels to network subscribers, buffer of handoff calls when

channels are not a vailable and allocation o f network facilities to b uffer new calls ba sed on the

frequency of channel demand. Finally, the performance of the new model is compared with that of the

non-combined (existing) model.

2. Analysis of Existing GSM Network Model

In a typical GSM network, the major determinant of system performance is always tied to traffic flow.

In this analysis, one of the basic assu mptions is that it is im practicable for all subscri bers to be

connected at the same time due to the resources shared among the subscribers. The under-lining issue is

the probability that a system will be congested (busy) and will be unable to serve a potential subscriber

[19]. Fig. 1 shows the g eneral architecture of GS M network while fig. 2 depicts schematic

representation of a GSM service centre.


IJOART

IJoART


Fig 1 General Architecture of GSM Network

Fig 2 Schematic Representation of a GSM Network Service Centre

It is believed that a GSM network system has a pool of limited resources (servers). A subscriber arrives

with intention of using one of the ser vers for a period of tim e (service time). If any of the servers is

available, it is ‘hel d /occupied’ by th e arriving subscriber. If none is a vailable, the arri ving call is

‘blocked’. Hence, the entire (offe red) traffic load of the syste m is then split into; served load and

blocked load. The qu estion remains, ‘how much of the offered load is served and how much is

blocked’? (Congestion concept)

Offered load ‘A’ =1

Where = Average call arrival rate, = service time, 1/ = Average holding time

3. The Combined Model

In this Co mbined Model, which is the proposed new model, Call A dmission Control ( CAC) and

Network Resource Allocation are the key issues. CAC deter mines the condition for ac cepting or

rejecting a new call based on the a vailability of sufficient networ k resources to guarante e the QoS

parameters without affectin g the existing calls. The major call-level qualities of service param eters

based on cellular telephone concept are: ne w call block ing and handoff call block ing probabilities.


IJOART

IJoART


However, for this new model design, instead of blocking such call, a buffer is introduced w hich stores

the call until a channel is available to trans mit the call. This recommended model has a pri oritization

scheduled to maintain un-interrupted communication during emergency. It also provides a location for

temporary memory to cater for handoff and incessant frequent callers within a specified period. It gives

them a higher priority over a new entrant call. This is a unique innov ation for this research. This model

should be implemented at every base station. The advantages of this operation are:

It gives a priori ty to hi ghly essential duties calls that need i mmediate attention; this w ill thereby

forestall any casualties that may occur if such attention is not given.

It gives priority to the most denied calls to grab the channel when they appear within a specified

time.

It does not allow any call to occupy the channel more than necessary when there are calls waiting to

grab the channel.

It does n ot pre-empt the subs criber if there i s no call waitin g and allows d ynamic allocation of

channel when there is equal priority calls.

3.1 The Combined Model Algorithm

Start()

Initiate_call()

Add_call_to_Queue()

Add caller ID

if((counter <= n)&&(callerPriority == higher)){

Assign_Channel()

}else{

if (Call == Handoff){ // hand off start

if (Channel == Free){


IJOART

IJoART


Assign_Channel

}else{

if(Buffer == Full){

End_Call

}else{

Buffer_Call_to_M1()

}

}

}else{ // hand off end

if (Call == Waiting){

Get_Caller_With_Highest_Waiting_Time()

}else{

if (Channel == Free){

Assign_Channel()

}else{

Buffer_Call()

}

}

}

}


IJOART

IJoART


3.2 Characterization of Combined Model

A cautious observation of this model reveals that this Combined Model has some distinct features when

compared with existing model.

There are three basic features found in this Com bined Model that accounts for its distinc tion. These

features include

Call prioritization technique.

Call buffer technique for handoff call.

Call waiting / frequently recent call allocation technique.

3.3 System (Combined Model) Testing

The three basic methods for testing and evaluation of GSM network performance include

Drive Test

Customer Complaint / Subscribers Feedback

Traffic Observation Report/Networ k Statistics from Network Operation and Maintenance Centre

(OMC) Report.

The method adopted for data collection in th is research is driv e test usin g the well known software

called the TEMS (Ericsson Test Mobile System), and fig. 4 is the setup for the drive test.


IJOART

IJoART


Fig 4 Block diagram of Drive Test Componenets Interconnectivity

The drive test is used because firstly it is a powerful tool for the radio frequenc y (RF) a nalysis and

problem solving, secondly, the scanner to ol used in drive test is a very good tool for detectin g

interfering signals, and finally, the dri ve test gives the exact geo graphical location for each sample

through the connected GPS receiver.

4. Result

The post processing was done via the assistance of a major GSM network service provider for both the

combined (C-M) and existing or non-combined model (E-M), considering the parameters such as Call

Set-Up Success Rate (CSSR), Call Drop Rate (CDR), Call Completion Success Rate (CCSR) and

Traffic Channel Congestion Rate (TCHCR), from January to April, 2014 and the g raphical forms of

these results are represented as figures 5, 6, 7 and 8.


IJOART

IJoART


Fig 5 Call Set-Up Success Rate for Combined and Existing Models

Fig 6 Call Drop Rate for Combined and Existing Models


IJOART

IJoART


Fig 7 Call Completion Success Rate for Combined and Existing Models

Fig 8 Traffic Channel Congestion Rate Combined and Existing Models

NOTE: C-M represents Combined Model E-M represents Existing Model

5. Discussion of Results


IJOART

IJoART


Figure 5 shows the Call Set-Up Success Rate (CSSR) of both the combined and the existin g systems

during the test period. It can be seen that the parameter is higher in the co mbined model than in the

existing model and this implies that combined model is better with respect to CSSR. Figure 6 depicts

the Call Drop Rate (CDR) of the two syste ms during the test period. Since the parameter is less in the

combined model than the exi sting model the implication is that the co mbined model is b etter with

respect to CDR.

Figure 7 defines the Call Co mpletion Success Rate (CCSR) of both the co mbined and th e existing

systems during the test period. As the parameter is hi gher in the com bined model than that of the

existing model it means that the combined model is b etter than the existing model with respect to

CCSR. Figure 8 represents the Traffic Channel Congestion Rate (TCHCR) of the two systems during

the test period. The fact that the para meter is less in the combined model than in the exist ing model

shows that the combined model is better with respect to TCHCR. Since the co mbined model has better

values than the exiting model in all the four k ey performance indicators used in this performance

assessment, it is therefore clear that the combined model offers a better quality of service (QoS) than

the existing model.

6. Conclusion

It has be en analytically pr oven that we can improve (optimize) an ex isting GSM network using

different methodologies to offer re markable Quality of S ervice to the end users. Moreo ver, the issues

discussed here are quite helpful for the an alysis and perform ance evaluation of differ ent mobile

networks. The priority calls function of the proposed model can be particularly useful in security,

health and other related emergency matters.

Introduction of buffers in this design plays a vital role in ensuring that the handoff and waiting calls are

not deteriorated and the quality of ser vice is generally improved. It has been established that the

proposed model offers a very hi gh quality o f service and therefore recommended for use in global

systems for mobile networks and can also be adapted to other networks.


IJOART

IJoART


References

1. J.J. Popoola, I.O. Megbowen, and V.S.A. Adeloye, “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, 2009.

2. S.M. Redle, M.K. Weber, and M.W. Oliphant, “Introduction to GSM,” the Atech House Mobile Communication

Series, U.S.A. 1995, pp. 78.

3. B.M. Kuboye, “Optimization for Minimizing Congestion in Global System for Mobile Communication (GSM) in

Nigeria,” Journal Media and Communication Studies, vol. 2, no. 5, Pp 122-126, 2010.

4. R. Syski, “Introduction to Congestion Theory in Telephone system,” Elsvier Science Publishers B.V. 1986, pp. 258.

5. E.S. Mughele, W.A. Olatokun, and T. Adegbola, “Congestion Control Mechanisms and Patterns of Call Distribution

in GSM Telecommunication Networks: The Case of MTN Nigeria,” African Journal of Computing and ICT, vol. 4,

no. 3, pp. 29-42, 2012.

6. O. Kazunori, and K. Fumito, “Dynamic Channel Assignment in Cellular Mobile Radio Systems,” IEEE International

Symposium on Circuits and Systems, vol. 2, no. 1, pp. 938-941, 1991.

7. A.S. Adegoke, I.T. Babalola, and W.A. Balogun, “Performance Evaluation of GSM Mobile System in Nigeria,”

Pacific Journal of Science and Technology, vol 9, no. 2, pp. 436-441, 2008.

8. B. Jabbari, and S. Tekunay, “Handover and Channel Assignment in Mobile Cellular Networks,” IEEE

Communication Magazine, vol. 30. no.11, pp. 42-46, 1991.

9. R. Ramjee, D. Towsley, and R. Nagarajan, “On Optimal Cell Admission Control in Cellular Networks,” Wireless

Networks, vol. 3, no. 1, pp 1-5, 1997.

10. A.S. Acanpora, and M. Naghushineh, “Control and Quality-of-Service Provisioning in High Speed Microc ellular

Networks,” IEEE personal communications, vol. 1, no. 2, pp. 36-43, 1994.

11. T. Sirin, “A Measurement-Based Prioritization Scheme for Handovers in Mobile Cellular Network,” IEEE JSAC,

Vol. 10, pp. 1343-1350, 1992.

12. N. Nasser, and H Hussanein, “Combined Admission Control Algorithm and Bandwidth Adaptation Algorithm in

Multimedia Cellular Networks for QoS Provision,” IEEE Proceedings, pp. 1183-1186, 2004.

13. G. Karagiannis, “Scalability and Congestion Control in Broadband Intelligent and Mobile Networks,” A Ph.D

Dissertation submitted to the centre of Telematics and Information Technology, University of Twente, 2002.

14. T.O. Oyebisi, and O.A. Ojesanmi, “Development of Congestion Control Scheme for Wireless Mobile Network,”

Journal of Theoretical and Applied Information Technology, Pp. 965-972, 2008.

15. A.S. Adegoke, and I.T. Babalola, “Quality of GSM Telephone System in Nigeria. American Journal of Scientific and

Industrial Research, vol. 2, no. 5, pp. 707-712, 2011.

16. A.M. Alorape, A.T. Akinwale, and O. Falarunso, “A Combined Scheme for Controlling GSM Network Congestion,”

International Journal of Computer Application, vol. 14, no. 3, pp. 47-53, 2011.


IJOART

IJoART


17. E.S. Mughele, T. Adegbola, O.B. Longe, and R. Boateng, “Factor Analysis for G.S.M. Service Congestion – The

Case of MTN Nigeria,” Computing and Information System Journal, vol. 16, no. 1, pp. 19-30, 2012.

18. E.S. Mughele, and W.A. Olatokun, “Comparative Evaluation of GSM Quality Service of Network Performance in

Nigeria Telecommunication Industry,” Computing, Information Systems and Development Informatics Journal, vol.

3 no. 3 pp. 23-34, 2012.

19. 0. Awodele, S.A. Adebanjo, S.O. Okolie, and E.E. Onuiri “Comparative Evaluation Models for Cellular

Communications Networks,” Proceedings of Informing Science and IT Education Conference (InSITE), pp. 304-305,

2012.


IJOART

IJoART

Documents

Comparative Analysis of the Performances of …...Combined model, the perfor mance of the GSM networ k is extensively evaluated using the key performance indicators which include;