Abstract—This paper presents a method for detecting periodic behavior and burst durations in IP traffic observed on access links. Periodic behavior in traffic observed on the Internet is growing due to the increasing amount of services with video components. The method is based on active probing and serial correlation of the observed probe packet inter-arrival times on the receiver side. Verification of the method is done based on an experimental implementation, using both controlled lab services and live public services. The services used in the experiments are adaptive video streaming based on Microsoft Smooth Streaming and Netflix. The results show that the method provides accurate results for both the estimator of service period and the burst durations. The findings are similar both for the controlled lab experiments and the live network measurements. The amount of probe traffic required in order for the method to perform satisfactory is very moderate.

Index Terms—Periodic traffic patterns, traffic burst duration, active probing, serial correlations, adaptive video streaming, access links.

The authors are with the Norwegian Institute of Science and Technology, Department of Telematics, Trondheim, Norway (e-mail: {bjorn.villa, poul.heegaard}@item.ntnu.no).

Cite:Bjørn J. Villa and Poul E. Heegaard, "Detecting Period and Burst Durations in Video Streaming by Means of Active Probing," International Journal of Computer and Communication Engineering vol. 2, no. 4, pp. 460-467, 2013.