Measuring mobile broadband performance in Nigeria: 2G and 3G

Abstract

With increase in broadband penetration rate in Nigeria there is very little known customer-centric mobile broadband performance analysis in the country, despite the inherent advantages associated with performance monitoring to regulators, operators, content-developers, and most especially the customers. There exists an information gap and customers are curious to know the Quality of Service (QoS) offered them. This paper presents a host and crowdsourced based approach to mobile broadband performance metric measurement and evaluation. A mobile broadband performance measurement application (MBPerf) was developed using Java and Extensible Markup Language (XML) and installed on volunteers’ Android Smartphones to measure and collect data relating to 4 (four) QoS metrics – download and upload speeds, latency and DNS (Domain Name Service) lookup; and user data such as mobile phone information, network information, and location information. Measurements were taken for a period of 3 months within Akure and Ibadan metropolis from the 4 major MNOs’ (MNO-A, MNO-B, MNO-C and MNO-D) networks in Nigeria. Data was retrieved from he cloud, pre-processed, sorted and analysed using Microsoft Excel version 13 and SPSS (Statistical Package for the Social Sciences) Statistics 19. Findings reveal that 3G users are not getting the industry set speeds. They get about 10% below the lower limit of the benchmark (500 kilobits per second). However, 2G users get a better deal of about 61% above the lower limit of the benchmark (100 kilobits per second). It was inferred that network performance is highly unpredictable and variable during the day (between 8am and 5pm) but greatly improves at the early hours of the morning (between 12am to 6am) with a difference of about 69% between the peak and worst performance. The study indicates that performance deteriorates at peak times (between 7pm and 11pm). Lastly the DNS performance analysis suggests that the MNOs’ DNS servers operate effectively and do not add significant delay to end users’ queries. Keywords: Mobile broadband performance, Quality of Service (QoS), crowdsourcing, MBPerf application, hostbased