Quantifying and Improving Internet video-QoE



Benefits of using (proposed) source initated frame restoration (SIFR) to improve video quality on the Internet.
While the default-IP path continues to degrade (left), SIFR is able to restore perceptual quality (right).



Video-QoE degradation at the Internet "link" level

The capability of present day Internet links in delivering high perceptual quality streaming services, is not completely understood. Link level degradations caused by intra-domain routing policies and inter-ISP peering policies are hard to obtain, as ISPs often consider such information proprietary. Understanding link level degradations will enable us in designing future protocols, policies, and architectures to meet the rising multimedia demands.

As a first step, this project presents a trace driven study to understand QoE capabilities of present day Internet links using 51 diverse ISPs with a major presence in US, Europe and Asia-Pacific. We study their links from 38 vantage points in the Internet using both passive tracing and active probing for six days. We provide the first measurements of link level degradations of intra-ISP and inter-ISP peering links from a multimedia standpoint. Our results indicate that Internet routing policies are not optimized for delivering high perceptual quality streaming services. Our study offers surprising insights into intra domain traffic engineering, peering link loading, BGP and the inefficiencies of using AS-path lengths as a routing metric. We also provide comprehensive case studies for every scenario we investigate. Streaming services apart, our Internet measurement results can be used as an input to a variety of research problems.


Effects of Internet path selection on video-QoE

The second part of this project presents large scale Internet measurements to understand and improve the effects of Internet path selection on perceived video quality. We systematically study a large number of Internet paths between popular video destinations and clients to create an empirical understanding of location, persistence and recurrence of failures. We map these failures to perceptual quality by reconstructing video clips obtained from the trace to quantify both the perceptual degradations from these failures as well as the fraction of such failures that can be recovered.

We then investigate ways to recover from QoE degradation by choosing one-hop detour paths that preserve application specific policies. We seek simple, scalable path selection strategies {\em without} the need for background path monitoring or apriori path knowledge of any kind. To do this, we deployed five measurement overlays: one each in the US, Europe, Asia-Pacific, and two spread across the globe. We used these to stream IP-traces of a variety of clips between source-destination pairs while probing alternate paths for an entire week. Our results indicate that a source can recover from upto 90\% of the degradations by attempting to restore QoE with any five {\em randomly} chosen nodes in an overlay. We argue that our results are robust across datasets.

Finally, we design and implement a prototype packet forwarding module called source initiated frame restoration (SIFR). We deployed SIFR on PlanetLab nodes, and compared the performance of SIFR with the default Internet routing. We show that SIFR outperforms IP-path selection by providing higher on-screen perceptual quality.


Related Publications