Basic Characteristics and Arrival Curve Characterization for Two-way Traffic of Online Video Watching

Abstract

In order to better guarantee different quality of service requirements for different types of traffic, characteristics of each traffic type should be studied and modeled. In this paper, a measurement-based study is reported, which includes traffic data collection and processing. For traffic modeling, the concept of cumulative arrival process is adopted. In particular, the arrival curve model in network calculus is used for traffic characterization. Four types of traffic data are collected and analyzed as examples. The results for an online video watching application under three resolutions are discussed in detail. As a novel aspect of the study, the traffic of the application on both directions, i.e., the two-way traffic, is considered. For basic traffic characteristics, the traffic rates and the probability density functions of packet length and packet interval are analyzed. To characterize the traffic, the corresponding arrival curves are derived and discussed. The method adopted in this paper may also be applied to other traffic cases.