Measurement-Based Analysis of Mobile Traffic in LTE-A Systems

Abstract

With the development in mobile communication systems, the number of smart devices have been rapidly increasing in these days, subsequently resulting in the increase in mobile applications as well. Although the current wireless communication system uses a wider spectrum and a higher frequency than previous communication systems to accommodate more mobile traffic, the spectrum still remains limited. As the number of smart devices increases, the transmission data rate supportable for each device may decrease due to the limited spectrum bandwidth. One natural and simple method to solve this problem is to increase the spectrum bandwidth for mobile communications. However, according to a survey on frequency spectrum released by the Commonwealth Scientific and Industrial Research Organization (CSIRO), the demand for wireless communications in Australia is on track to reach ‘peak data’, outstripping the available radio frequency bandwidth spectrum by 2020. They warned such a risk assuming the current network speed and the current growth of data usage. To prevent this situation, it is important to understand the practical resource usage in current LTE-A systems, which may serve for the future deployment of mobile communication systems.

With such motivation, in this thesis we measure and observe mobile traffic of three service providers in Korea (SKT, KT, and LGU) from two perspectives: terminal and eNodeB. First, from a terminal perspective, we consider measurement scenarios for downloading streaming video at cell edges using Spiretech’s Diagnostic Monitor (DM) equipments. After pre-processing the collected data, we analyze the required data rate to support video streaming, spectral efficiency based on channel condition, and resource assignment pattern over time. Second, from the eNodeB’s perspective, we perform measurements in crowded places where traffic saturation is expected. We then analyze the distribution of key parameters in downlink, data rate, and spectral efficiency. In addition, we propose a mathematical model to analyze the traffic capacity of the LTE-A systems. Then, the accuracy of the traffic capacity model has been verified through comparing with the measurement data. It is expected that the proposed mathematical model could also serve to estimate the traffic capacity of future cellular systems.