Evaluation of User Side Quality of Service Measurement Techniques for Addis Ababa LTE Data Service
No Thumbnail Available
Date
2018-07
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
AAU
Abstract
Penetration of high-rate and innovative mobile data services has increased to the global mobile
industry such as social media, video streaming, and other multimedia services. To
accommodate such innovative mobile data services, Long Term Evolution (LTE) mobile
technology is widely deployed and operational in several countries including Ethiopia where
Ethio telecom has launched LTE data service in the capital Addis Ababa on March 2015.
To maintain and sustain satisfactory LTE data services, LTE quality of service (QoS) needs to
be measured and analyzed so that informed optimization and other decisions are made by
operators, regulators and users. Various QoS parameters of data services can be measured
using different network and user side measurement techniques. Drive test is the most common
user side QoS measurement method that is applied by mobile operators and recently
crowdsourcing techniques, where client applications are used to measure and collect QoS data
from users, are being applied.
In this thesis, we evaluate the performance of user side QoS measurement techniques for LTE
data services focusing on selected crowdsourcing techniques: OpenSignal and Speedtest. We
also present accuracy performance comparison between the crowdsourcing techniques and the
conventional drive test. The performance evaluation and comparison are performed using
measurement data that we have collected for selected ‘Tikur Anbesa’ area route of Addis
Ababa. Detail analysis of the measurement techniques based on measured data is provided
using Matlab. Furthermore, to understand LTE quality perception by various actors of the
mobile industry, at the beginning of the thesis, we undertook a survey for Addis Ababa LTE
network to the enterprise, end users, Internet Service Provider (ISP) and to the ISP regulator
Ministry of Communications and Information technology (MCIT). Results show that, there is a limited awareness to the available measurement techniques and
the observed LTE data service is below the expected average value, a summarized result of the
survey are discussed in chapter two Section 5.3.3 of this thesis work.
To access the quality of experience for LTE data service a downlink throughput and latency
key performance indicators (KPI) are measured. Through the analysis of the measurement
techniques: OpenSignal, SpeedTest and Nemo Handy for the KPI, we identify a significant
difference in latency between OpenSignal and Nemo Handy; on the contrary the difference in
latency between SpeedTest and Nemo Handy is minimal. The downlink throughput difference
between OpenSignal and Nemo Handy is minimal, whereas SpeedTest and Nemo Handy has
a major difference. We have quantified and evaluated the accuracy between the measurement
techniques using Root Mean Square Error (RMSE) metric. The key factors affecting the results
in addition to the test server location are discussed.
Description
Keywords
LTE, Crowd Sourcing, Drive Test, QoE, QoS