On Mon, 28 Oct 2024 at 19:22, Stuart Cheshire <cheshire=40apple.com@xxxxxxxxxxxxxx> wrote:
Please send comments to ippm@xxxxxxxx <mailto:ippm@xxxxxxxx>, or attend IPPM in Dublin to share your thoughts in person.
Stuart, thanks for pre-emptively identifying the most appropriate list in the first message! I agree this is a topic where it's valuable to flag it to the whole ietf@xxxxxxxx community.
For anyone wanting to respond, the link to subscribe to IPPM is https://mailman3.ietf.org/mailman3/lists/ippm@xxxxxxxx/.
While I'm here, I'll add my thoughts.
Is this a good predictor of video conferencing performance?
In networks where non-congestion packet loss is rare (i.e. most packets do arrive eventually) I agree that user experience is dominated by the latency of the most-delayed packets, not the median. I'm not sure which centile is optimal, but personally I would start with 99th and go upwards from there (i.e. 99.x percent). There are some people in BT R&D who've looked in this space so I'll reach out to them.
You might also find this informative:
Jia He, Mostafa Ammar, Ellen Zegura, Emir Halepovic, and Theo Karagioules. 2024. QoE Metrics for Interactivity in Video Conferencing Applications: Definition and Evaluation Methodology. In Proceedings of the 15th ACM Multimedia Systems Conference (MMSys '24). Association for Computing Machinery, New York, NY, USA, 178–189. https://doi.org/10.1145/3625468.3647622
Thanks
Chris
