Speakers:
Deep Medhi, University of Missouri-Kansas City

Measuring Quality of Experience (QoE) with YouTube from users accessing from Residential ISPs is often a difficult task. First, there is the problem of a good data collection mechanism, second is the issue of being able to obtain precise information on metrics that impact QoE such as video start time, number of interruptions, duration of interruptions, etc. Due to YouTube's content distribution architecture, it is also important to be able to obtain information on how the location from where a video is served impacts QoE metrics for end users.

We have developed Pytomo, a tool that emulates YouTube viewing experience by end users, through which we can collect a number of QoE metrics along with information on CDN. The tool is made available in the public domain at [ Ссылка ]

Using this tool, we have collected data from end users accessing YouTube from three different residential ISPs. In this talk, we will present observed QoE along with impact due to CDN. We will also discuss how such a tool can be useful to ISPs ( (it has been deployed by one ISP).

This is a joint work by Parikshit Juluri*, Louis Plissonneau+, Yong Zeng*, Deep Medhi* (*University of Missouri-Kansas City, USA; +Orange Labs, France)

See more at: [ Ссылка ]

свернуть