Re: [EToSat] Interop runner with satellite links

[Joerg Deutschmann <[e-mail]>]

Dear CJ,

stepping in for Sebastian, who has earned a break after finishing his 
thesis.
https://interop.cs7.tf.fau.de/thesis.pdf


> Sec. IV,D, 5: Interesting to know "picoquic" does speculative retransmission. As you argue, this may not always help. Did you confirm with the author
No, we have not discussed results with any author, simply because the 
number of implementations is too high to understand each implementation 
in detail. I added Christian Huitema in Cc, maybe he wants to comment on 
this one.

> The results show loss-based CC does not perform well compared to BBR
Yes, but as mentioned in the paper, we are not sure about the CC of each 
and every implementation.

> Production software performs well more like picoquic or not? How big difference is between production vs these implementations in the test?
Good question, we don't know, should be noted as future work. Sebastian 
only has a very short section "Comparison with Results of Other 
Research" in his thesis. We assume that Docker does not have a big 
influence on the performance, but did not verify that. Also chrome not 
being part of the tests is unfortunate.

> Any Time-Offset graphs for EUTELSAT case?
Yes, on the https://interop.cs7.tf.fau.de website select "Run: 
real_2021-11-15" and then "Plots for eutelsat (EUT)". We plan further 
runs in the future.

> Research overview page: additional column on emulated or real Sat link will be helpful
Actually that column already exists, 5th column from the right.
Motivation for putting the overview on github was to allow pull 
requests, but visualization and markdown editing for such a large table 
is not optimal with github...


Thanks,
Joerg