Analysing the characteristics of VoIP traffic
In this study, the characteristics of VoIP traffic in a deployed Cisco VoIP phone system and a SIP based soft phone system are analysed. Traffic was captured in a soft phone system, through which elementary understanding about a VoIP system was obtained and experimental setup was validated. An advanced experiment was performed in a deployed Cisco VoIP system in the department of Computer Science at the University of Saskatchewan. Three months of traffic trace was collected beginning October 2006, recording address and protocol information for every packet sent and received on the Cisco VoIP network. The trace was analysed to find out the features of Cisco VoIP system and the findings were presented.This work appears to be one of the first real deployment studies of VoIP that does not rely on artificial traffic. The experimental data provided in this study is useful for design and modeling of such systems, from which more useful predictive models can be generated. The analysis method used in this research can be used for developing synthetic workload models. A clear understanding of usage patterns in a real VoIP network is important for network deployment and potential network activities such as integration, optimizations or expansion. The major factors affecting VoIP quality such as delay, jitter and loss were also measured and simulated in this study, which will be helpful in an advanced VoIP quality study. A traffic generator was developed to generate various simulated VoIP traffic. The data used to provide the traffic model parameters was chosen from peak traffic periods in the captured data from University of Saskatchewan deployment. By utilizing the Traffic Trace function in ns2, the simulated VoIP traffic was fed into ns2, and delay, jitter and packet loss were calculated for different scenarios. Two simulation experiments were performed. The first experiment simulated the traffic of multiple calls running on a backbone link. The second experiment simulated a real network environment with different traffic load patterns. It is significant for network expansion and integration.
DegreeMaster of Science (M.Sc.)
SupervisorMakaroff, Dwight; Eager, Derek L.
CommitteeTeng, Hsiang-Yung (Daniel); Osgood, Nathaniel; Jamali, Nadeem