A Statistical Framework to Monitor the Quality of Service in Mobile Networks

2 Pages Posted: 31 Mar 2017

See all articles by Tania Villa

Tania Villa

Instituto Federal de Telecomunicaciones

Nimbe Ewald


Date Written: March 15, 2017


In Mexico, the mobile network operators have reached penetration rates of about ninety percent of the population. However, the offered Quality of Service (QoS) still varies drastically between geographical areas due to, among others, the fact that the infrastructure deployed, for instance in rural areas, is not the newest access technology. Moreover, there is still a need to enhance the capacity of current mobile networks to satisfy the service demands of current and future users and applications.

In countries with no effective competition, this can be fostered by the telecommunications regulator whose role would be to define a set of metrics to help the enforcement of minimum standards for quality of telecommunication services. Furthermore, providing information produced by continuously monitoring QoS will provide a valuable source of data that can be used to empower the final users by keeping them informed about the QoS offered by operators enabling them to take decisions accordingly.

In this context, the recent Resolution 95 (2016) of the World Telecommunication Standardization Assembly resolves that the ITU Telecommunication Standardization Sector "provide references that assist developing and least developed countries in establishing a national quality measurement framework suitable to perform QoS and QoE measurement". Furthermore, it instructs study groups of the ITU Telecommunication Standardization Sector, among others, "to elaborate Recommendations providing guidance to regulators in regard to defining strategies and testing methodologies to monitor and measure QoS and QoE" and "to study scenarios, measurement strategies and testing tools to be adopted by regulators and operators".

Given this and the fact that there are not yet normalised technical specifications or recommendations targeted for regulators, in this paper, we propose a system of metrics to assess the mobile telecommunication services offered in Mexico (voice, short message service (SMS) and data transfer) as well as a methodology to monitor the QoS at a national level and to measure the proposed metrics.

We develop a two-step statistical modeling approach using a stratified random sampling in the first step to select the geographical locations to be measured and a simple random sampling in the second step to determine the sample size for each service to be tested. We describe the procedure to construct the strata by selecting non-overlapping groups from the geographical regions in the country. The idea behind using stratification is to produce a smaller bound on the error of estimation than would be produced by a simple random sample of the same size alone. Instead, we use stratification combined with simple random sampling in each stratum to estimate national values for each QoS metric defined.

A theoretical example of implementation of the testing methodology at national level in Mexico is also presented to show its feasibility considering factors such as access technology, current service coverage and working days needed to perform the measurements.

Finally, we outline a set of recommendations that can be customized by any regulator to measure the performance and QoS at local and national level.

Keywords: quality of service, sampling methods, mobile networks, regulation, policy enforcement

Suggested Citation

Villa, Tania and Ewald, Nimbe, A Statistical Framework to Monitor the Quality of Service in Mobile Networks (March 15, 2017). Available at SSRN: https://ssrn.com/abstract=2944305

Tania Villa (Contact Author)

Instituto Federal de Telecomunicaciones ( email )

Insurgentes Sur #1143, Col. Nochebuena
Delegación Benito Juárez
Mexico City, 03720

Nimbe Ewald

Independent ( email )

No Address Available
United States

Here is the Coronavirus
related research on SSRN

Paper statistics

Abstract Views
PlumX Metrics