2008 · 20 citations · 5 references
ComputingCommunicationComputer SystemsComputer NetworksDigital CommunicationPerformance MeasurementCommunication NetworksCommunication SystemsJitter AnalysisNetwork Performance AnalysisEmpirical EvaluationHurst Exponent EstimationJitterJitter EstimationNetwork JitterPacket Loss AnalysisSelf-similar ProcessesSelf-similar Traffic
In this paper, as result of a thorough self-similar analysis we found self-similar characteristics in different sets of VoIP Jitter traces, with short range and long range statistical dependence (i.e. LRD and SRD). We used the E-model's R factor for estimating the QoS perceived by the end user, and we established the following empirical relationships: when the values of the estimated Hurst parameter were lower than 0.5 (SRD) the analyzed performance metrics, i.e. average packet loss rate and average Jitter, maintained stable values and acceptable QoS levels. While the values of the estimated Hurst parameter were greater than 0.5 (LRD) the variability of the analyzed performance metrics were increased and the QoS levels diminished. The other empirical relationship shown in this article refers to the payload size for each CODEC type used, the larger the payload size was, the higher the average Jitter became.
5
On the self-similar nature of Ethernet traffic
W.E. Leland, Murad S. Taqqu, Walter Willinger et al. · 1993 · 1.1K citations
Bandwidth estimation: metrics, measurement techniques, and tools
R. Prosad, C. Davrolis, M. A. Murray et al. · IEEE Network · 2003 · 732 citations