Some mobile applications can generate data traffic on the network which we can observed and we can determine the data profile pattern characteristic for the application. From this pattern, we can conclude the behavior the user or the actual characteristics the used network or application, because this all have effect to the packet level characteristics of the traffic. The quality is gradable with the help of key performance indicators of each profile pattern. This helps us to detect the signs of the quality drops of the network and the services. The subject of my dissertation is the popular image sharing application the Instagram, which helps me to present used detection methods.
The purpose of the dissertation is to implement an automated control system that is based on the data traffic profile of the application, which is capable of the dynamic managing if the dynamic resources. The whole system can be divided into two major parts, the purpose of first component is to capture the generated date traffic on the network and to detect the application and different use cases based on captured data. In my dissertation, I will present these detection methods that helps to create data traffic profile of the that helps to configure the control parameters. Furthermore, I demonstrate the efficiency of the detections methods trough the Instagram. The second part is the control of resources based on the data processed by the monitoring components. In the task the measure of the user experience is a key role. If this indicator goes below a certain level than the system component encroaches in the process. We worked on the task in a group of three, which included the examination of various applications and the implementation of the whole system. I designed and implemented the control component in C and C++ languages, detailing of it’s functioning.