Network Traffic Analysis for online meeting characterization: Estimation of the type of meeting exploiting features extracted from rtp packets headers.

The Internet has pervasively transformed business activities. Forging this principle ahead, online meetings are becoming the major solutions for business communication when in-presence meetings are not feasible or as effective solution to costs savings. QoS/QoE network management policies are needed to ensure a quality of experience of online meeting comparable to in-presence meeting experience. In this context, it is important for network service providers to understand which type of meeting is taking place and how many participants are involved in the communication. The purpose of this thesis is to address this problem and propose a working algorithm to effectively characterize online meetings exploiting network traffic analysis techniques in order to define network management policies transparently to customers and applications. The thesis provides a method to extract information related to the maximum and mean number of participants throughout all the meeting and to the type of meeting. In addition, it also provides a method to estimate the type of meeting while the meeting is in progress. This approach can make up the first step of a QoS/QoE management policy to improve the overall meeting quality of experience.