Analysis and classification of VoIP traffic for mobile devices

Show full item record



Permalink

http://urn.fi/URN:NBN:fi-fe201804208662
Title: Analysis and classification of VoIP traffic for mobile devices
Author: Haris, Muhammad
Contributor: University of Helsinki, Faculty of Science, Department of Computer Science
Publisher: Helsingin yliopisto
Date: 2018
Language: eng
URI: http://urn.fi/URN:NBN:fi-fe201804208662
http://hdl.handle.net/10138/273500
Thesis level: master's thesis
Abstract: The number of applications that are used for voice and video calls is growing day by day. The basic purpose of all of these applications is to provide better call services to their customers. These applications differ in their underlying communication protocols and encoding techniques. These variations in protocols are introduced to enhance the user experience, improve the performance of the VoIP applications and minimize the network delay between the end users. Thus, to make the call quality better, there is a lot of research done to analyze and detect VoIP traffic. There are two basic goals of this thesis: 1) Analysis of the VoIP traffic of five famous VoIP applications (Skype, Viber, WhatsApp, Facebook messenger, and IMO). 2) Classification of VoIP traffic based on the analysis in the first step. We adopted flow-based analysis technique for the analysis and classification of VoIP traffic. For the first step, we analyzed the three flow features (packet rate, Inter-packet gap, and packet size) of the above-chosen VoIP applications. In addition, we also presented a detailed explanation of the three factors (the access network, the underlying operating system, and the geographic distance between the caller and the callee) which can affect distribution of the flows features. To realize our second goal, we present an analysis algorithm that uses moving weighted average (EWMA) and standard deviation (EWMSD). Our analysis algorithm is based on the statistical distribution of EWMA/EWMSD for each of the selected flow features mentioned above. The advantage of using EWMA/EWMSD statistics for classification is that they are not dependent on network protocols. Thus, a generic classification approach can be applied to all of the three flow features described above. Finally, based on our analysis, we introduce a classifier algorithm and its variants to classify a VoIP call as voice or video call. The classifier algorithm monitors EWMA/EWMSD statistics for the flow features and classifies the content of VoIP call accordingly. Our results explain two important things: 1) Each of the above-mentioned flow features can be used with EWMA/EWMSD statistics for VoIP traffic classification. 2) The use of Inter- packet gap feature with EWMA/EWMSD statistics gives highly accurate results as compared to the Packet size and Packet rate.


Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show full item record