Traffic Data Collection and Analysis for Profiling Mobile Applications and Improving their QoE

OData support
Supervisor:
Dr. Varga Pál
Department of Telecommunications and Media Informatics

By the year 2017, the Internet became a part of our everyday life, especially the mobile data usage has grown significantly in the past few years. In this thesis I am writing about a mobile application analysis about Tinder.

The measurement of these applications is needed for various reasons. Partly because we can make feedbacks to the application developers, as well we can map a user’s online behavior through the network model. Almost every application’s every function has a kind of a footprint, that shows a specific interaction’s online appearance. If we see this „image” (and we know the endpoints of the communication: the servers and clients), we could tell what a user is doing (just browsing on the Internet, watching videos, etc.).

By making these profiles we can not just analyse, but we have the chance to make these applications better. This improvement can be inside of an app (when we avoid encryptions on specific informations to speed up the operation) or - during the usage – to optimize the network traffic’s QoE gauges (if there’s a nonoptimal route between the server and the client).

In the past year there was made a network analyzer program that can be parameterized. From the domain names that Tinder uses and if the data is encrypted from the parameters of the TLS handshake and from the transfer speed it makes an indication of the user’s behavior.

As a normal user we only care about the latency and the bandwidth when we make opinions about an application. After making the profiles we can make different performance indicators, then helping the improvements of the SDNs (Software Defined Networks).

Downloads

Please sign in to download the files of this thesis.