Congestion Control for WebRTC Services

Abstract

Multimedia applications are increasing in popularity and using a big part of the Internet traffic. Web Real-Time Communication (WebRTC) is new technology which allows peer-to-peer communication in the browsers without any extra plugins. The focus of this report is to determine the importance of congestion control for the WebRTC-services.

In this thesis, I developed my own WebRTC-application. The developing process contained research about features, protocols, and technologies used. The final service was fully developed with audio and video features for multimedia conversations. In addition, instant messaging was added for assurance in case of communication problems.

Further, the developed WebRTC service were to conduct experiments. The Experiments conducted where divided into two phases. First phase tested the WebRTc service itself, to ensure the service was working properly. Second phase consisted of 10 participants using the developed service and give session feedback. I collected both session statistic data and giving feedback from each session.

The experiments focused on the Quality of Service (QoS) and the user perceived Quality of Experience (QoE), by looking at sent and received bits and packets, packet loss rate and jitter values. I got an indication of the QoS and if high values had an affect on how the user experienced the session. Results show that it was the audio interruptions were most heavily influenced by poor QoS.

Congestion control is a mechanism needed for transportation of data across the Internet, to promote fair usage and prevent congestion collapse. After looking at results from the experiments, a congestion control may be needed. I have evaluated two WebRTC congestion controls, Google Congestion Control (GCC) and Network Assisted Dynamic Adaption (NADA), to fully understand how they operate.

The findings from experiments and the evaluation indicate a need for a congestion control, and both GCC and NADA are found to be appropriate congestion controllers.

Keywords WebRTC, Quality of Service, Quality of Experience, congestion control, GCC, NADA, experiments.