Transport Protocol Influences on YouTube Videostreaming QoE

Video streaming currently dominates global Internet traffic and will be of even increasing importance in the future. In this paper we assess the impact of the underlying transport protocol on the user perceived quality for video streaming using YouTube as example. In particular, we investigate whether UDP or TCP fits better for Video-on-Demand delivery from the end user’s perspective, when the video is transmitted over a bottleneck link. For UDP based streaming, the bottleneck link results in spatial and temporal video artifacts, decreasing the video quality. In contrast, in the case of TCP based streaming, the displayed content itself is not disturbed but playback suffers from stalling due to rebufferung. Due to the lack of existing Quality of Experience (QoE) models for online video services that are based on TCP-streaming, we propose a generic subjective QoE assessment methodology for multimedia applications (like online video) that is based on crowdsourcing a highly cost-efficient, fast and flexible way of conducting user experiments. We demonstrate how our approach successfully leverages the inherent strengths of crowdsourcing while addressing critical aspects such as the reliability of the experimental data obtained. As a result, we present a dedicated QoE model for YouTube that takes into account the key influence factors (such as stalling events caused by network bottlenecks) that shape quality perception of this service. The results of subjective user studies for both scenarios (UDP based on related work, TCP based on own studies) are analyzed in order to assess the transport protocol influences on Quality of Experience of YouTube. To this end, applicationlevel measurements are conducted for YouTube streaming over a network bottleneck in order to develop models for realistic stalling patterns. Furthermore, mapping functions are derived that accurately describe the relationship between networklevel impairments and QoE for both protocols.