Skip to Main Content
Measurement of Internet Protocol Packet Delay Variation (IPDV) is a relevant issue in real-time video-streaming services. To assess the Quality of Service (QoS) to the end user, IPDV measures are commonly compared to conventional thresholds, whose values depend on the application requirements. However, exceeding these thresholds only for short time intervals and with relatively low repetition rates does not generally imply a significant decrease of the quality perceived by the users. Vice versa, IPDV values not exceeding the admitted threshold but very close to this value for long time intervals and characterized by relatively high repetition rates can significantly affect the quality perceived by the users. Consequently, studying the correlation between IPDV measures and some performance metrics directly related to the quality perceived by the end user is a very useful and challenging issue. Literature concerning with IPDV evaluation does not provide relevant information about the measurement procedure, such as measurement rate, way of collecting the obtained results in a single estimator, time interval during which the estimator is evaluated, and so on. The absence of these important aspects makes the comparison between IPDV measures and the aforementioned thresholds very critical, thus making even more ambiguous the evaluation of the QoS to the end user. In this framework, the aim of this paper is twofold. From one side, it aims at experimentally assessing and correlating the values that some performance metrics assume at different layers of a wide area network (WAN) supporting video-streaming applications for several values of IPDV; from the other side, it aims at deducing helpful and practical hints for designers and technicians, in order to efficiently assess and enhance the performance of a WAN supporting video streaming, in some suitable setup conditions affected by IPDV. The final goal of the research activity is to identify and fine tune an IPDV estimator- able to reliably estimate and detect the objective service degradation from the user perspective. The obtained estimator should be sensitive to service degradation that disregards the user's expectations, and at the same time, it should be insensitive to disturbs that do not cause a perceptible service degradation.