Problems of congestion in mobile networks

on   /   in Technology 12:00 am   /   Comments

By Augustine Odinma

I read with a keen interest the article of Dr Balarabe Mohammed Sani, NCC Director for Technical Standards and Network Integrity, which Vanguard carried on June 1, 2011. It was interesting that the Director began by acknowledging that it is generally agreed that the operators’ networks have serious issues of quality of services.

First, it is important that the Director is told that a similar research was conducted in London on 2nd, 4th and 7th of February, 2004 on two Operators and the highest recorded percentage failures was 0.001% for one of the operators and none for the other.

The problem of congestion in Nigeria is not a personal thing; it is a national calamity, which some of the government officials in position of authority fail to appreciate.

This is where one of the hindrances to the solution lies. Peter Drucker, a well known management doyen, said that ‘an identified problem is half solved’.

I am sure that the Director carries two or more phones like some of us who wish to be able to communicate at all times. Had he actually appreciated the problems mobile users are experiencing, his article would have been able to tell the readers what NCC is doing to readdress the problem, the progress they are making and how soon consumers should expect to feel the impact of their efforts.

Alternatively, he would have called a round table of experts in the field to discuss the problem of poor mobile phone quality, because it is a national problem that has been lingering since GSM was first launched in Nigeria. This is not the time for complacency or self denial, but a time to face reality.

I will briefly discuss our research and associated outcomes, I will then discuss the Director’s erroneous assertions. Finally, I will make observations and conclude with recommendations.

Research into Congestion in the Network

The research that we carried out and published in Vanguard Newspaper of 17th and 24th of May 2011 proved four things:

1) that mobile operators’ networks are enormously congested and that congestion is on the rise between 2005 and 2011;

2) that there is inconsistent and variable services delivery by operators from one big city to another, with the smaller cities on the receiving end;

3) that the numerous announced or displayed error messages do not represent or reflect the actual problem encountered in the mobile networks and

4) that registration of SIM cards, though plausible, has introduced another layer of hardships for mobile customers in smaller cities.

It was clear what our research was all about and equally clear is the definition of the research scenario. We quote from the paper under reference:

“The two test phones for each operator are place uninhibited on a desk in the test room. Both phones are therefore registered with the same Base station subsystem.

So, near perfect results were expected because this test scenario was the Least Common Denominator of fault expectancy. Results from test calls between two phones in different base stations or switching centres should produce more challenging outcomes.

Rural area calls would even be more challenging. So, it should be noted that based on the test scenario chosen, a call failure of more than 1% represents a serious congested network and it is therefore unacceptable.”

The articles are not about exhaustive quality of service and we never intended it to be so. The Director’s article was as though that was the case, but this tantamount to a redefine of the premise and consequently the ensuing conclusions.

Moreover, the articles were not intended for a technical audience. Let me first, for the benefits of the users, discuss congestions and thus show that the Director erred in his analysis.

Congestion is the common cause of “drop calls” or a phone user’s inability to make through-connected calls. A drop call is one of over a hundred (100) quality of service indicators.

Our research is only interested in that one indicator. Drop calls can occur when you seek to connect to another phone number – during call setup or during conversation and we are only interested in the former.

So, we defined and carried out a research in a test scenario, such that no drop call was expected. The equipment the Director would have preferred us to use for the test is one that would cost several millions of Naira and would give extraneous results, some of which would be irrelevant to proving congestion and outside what we set out to do.

Let us now consider some of his calculations and assertions, which show that he did not understand the research we carried out.

Dr Sani’s Erroneous Assertions

Dr Sani erroneously said that we employed wrong methodology of measurement and contended that it would be impracticable to have up to12,000 samples per day. We never claimed to have done that, but his calculations were in error.

The 12,000 samples should have been divided between the three tests locations (12,000/3 = 4,000) and further into two individual researchers carrying out the test (4,000/2) in each location and that would work out to be an average of 2,000 samples per student per day.

So, it is either that the Director intentionally gave the erroneous figure or he did not take the time to understand the research methodology.

Moreover, he alluded to the fact that we did not consider mobility management functions of the mobile network. We do not have to! Looking at the premise of our research, it should be clear that we did not and do not have to do so because that complicates the research.

Moreover, that would yield more drop calls. We said that we have chosen a scenario that should be error-free and that a drop call of 1% represents a serious congestion because of the scenario chosen. A mobile user can be stationary or can be mobile.

The user that is moving would yield more challenging results than one who is stationary, and consequently yield more drop calls. We were going for a zero-error scenario and that is what we did.

The Director postulates that “a stationary user may experience congestions because he is hooked to only one cell called serving cell that has weaker signal or traffic congestions at some instant”.

This is preposterous. Our test scenario said that we have chosen a place where there is full signal and he wants to create a make-believe weak cell.

The Director said that I did not consider bouncing busy hour. Why should I or why should the users of mobile phones care about bouncing busy hour? That is for operators to do and ensure that there is no intermittent insurgence of users moving to a particular area at some periods of a day as to cause congestion during that period.

What we are interested in is whether there was congestion or not, period! The cause of the congestion is not the subject of our research. The Director tried to redefine the premise of our research, but his analysis thereto are irrelevant, misplaced and technically mismatched.

He ended with a paragraph in conventional testing techniques of quality of service (QoS) measurement of mobile networks. Conventional methods are not the only way for experts to acquire results! He knows that we would not be able to financially acquire the equipments for conventional testing and also the data logs between operators for further analyses.

But, do we really need these to forecast that the networks are heavily congested? NO! The Director even contradicts himself because NCC only needs to monitor few key performance indicators of the operators; some of the other indicators he described are academic to the reader and are only necessary for the operators to ensure quality assurance.

Observations

The reason for publishing the articles was to show that we have serious congestions in our networks and that has not been disputed by the NCC Director in his article, where he tries to consider exhaustive QoS parameters.

He said towards the end of his paper that “in addition to hundreds of QoS statistical parameters, the equipment can perfectly evaluate speech quality of connected calls.”

The test for congestion does not need 100 indicators to describe it and NCC would only require the operators to meet few key performance indicators. So, one wonders why the Director expects our research to digress into areas that will either produce more drop calls or are not necessary to conclude that the networks are congested.

I do not know the Director and our article is not about any individual or NCC; It is about a major national problem. Our articles are not about showing our pedantic prowess; it is equally not about showing our knowledge in telecommunications. The articles we wrote in the news media were generally about poor operators’ services.

Conclusions and Recommendation

I am still yet to fathom why the NCC Director of Technical Standards and Network Integrity attacked the articles we published on the 19th and 25th of May 2011 about operators poor quality of services.

The primary job of a regulator is to protect consumers and not the operators. He challenged the test for congestion, but made no attempt to address the other outcomes of the research, one of which is that the operators’ services are inconsistent from one city to another, with users in the smaller cities suffering the most.

The other outcome was that the SIM card registration has introduced another layer of hardship for mobile users in smaller cities such as Yola or Yobe. The final outcome was that mobile operators announce or display many error messages that are incongruent with the real problem, probably to mast or camouflage the enormous problem of congestion in the networks.

The research we recently conducted in Lagos, Abuja and Yola was first conducted in London in February 2004 on two operators for three days and we observed a maximum failure of 0.001% for one of the operators.

In 2005, a similar research was conducted on four operators in Nigeria. So, one of the articles compared congestion observed for 2005 with the latest observation for 2011.

Furthermore, the articles are not about any individual or NCC, but about what has become a national problem. The Director agreed in a sentence that there are serious issues with quality of service, but took over hundred sentences to carry out his defence.

Prof. A C Odinma is a Professor of Telecommunications and IT, American University of Nigeria. He was formerly, Head of Mobility Network Solutions for Europe, Middle East & Africa, Lucent Technologies Bell Labs. Moreover, formerly, Senior Technical Manager, AT&T Labs, NJ, USA.

    Print       Email