DocumentCode :
124311
Title :
Modelling download throughput of LTE networks
Author :
Cainey, Joe ; Gill, Brijesh ; Johnston, Samuel ; Robinson, John ; Westwood, Sam
Author_Institution :
OpenSignal, London, UK
fYear :
2014
fDate :
8-11 Sept. 2014
Firstpage :
623
Lastpage :
628
Abstract :
We report the initial findings of an investigation that attempts to model download throughput of LTE networks in the US using the results of 130,296 speedtests performed by users of the OpenSignal mobile application on Android devices. We have developed a simple model which regresses LTE download throughput on a number of independent variables. We find that signal quality and strength, temporal and network operator factors are all essential, and also find an interesting device-specific dependence.We show that none of the LTE-specific signal metrics is sufficient alone to determine download throughput, although once combined they collectively explain an appreciable fraction of the variability in the model. We obtain time-of-day results which we believe act as a proxy for network congestion, and are able to quantify the relative performance of the four largest US cellular operators.
Keywords :
Long Term Evolution; cellular radio; smart phones; Android devices; LTE download throughput; LTE networks; LTE-specific signal metrics; OpenSignal mobile application; US cellular operators; device-specific dependence; network congestion; network operator factors; signal quality; signal strength; speed tests; Androids; Humanoid robots; Mobile communication; Mobile computing; Performance evaluation; Throughput;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Local Computer Networks Workshops (LCN Workshops), 2014 IEEE 39th Conference on
Conference_Location :
Edmonton, AB
Print_ISBN :
978-1-4799-3782-0
Type :
conf
DOI :
10.1109/LCNW.2014.6927712
Filename :
6927712
Link To Document :
بازگشت