Re: [tsvwg] Gorry Fairhurst Individual thoughts on choosing whether/how to advance ECN work.

[Jonathan Morton <[e-mail]>]

> On 24 May, 2020, at 2:53 am, Bob Briscoe <[e-mail]> wrote:
> 
> * cloud-rendered interactive control (like this: https://riteproject.eu/dctth/#1511dispatchwg )

The demo linked there is a comparison between DCTCP and a dumb FIFO - of course the latter is going to lose.  But plain ECN (or SCE) is already capable of getting queue delay below one frame of video, so this is not a compelling case for L4S.

> * remote presence,

See above.  But also note that for properly functioning VR, developers have had trouble keeping the input-to-display latency sufficiently low to avoid motion sickness without ever leaving the single machine that both input and display are physically attached to.  Doing that over any sort of network is going to be extra hard.

For this particular category of applications, the total latency of the network path will govern the experience, not so much the variability as long as the latter is reasonably small.  If you're talking with someone who's 160ms of light-time plus video codec latency away, an extra 5ms will not be noticed.  If they're just the other side of the campus, a total of 5ms will not be noticed.  Do you see the pattern?

> * interactive real-time remote control or supervisory monitoring of cameras, drones, machinery (oil rigs, power plants, machine shops, cranes, ), medical procedures, using (high res.) video to monitor what is being controlled.

Are you seriously suggesting running realtime, safety-critical processes over a best-effort Internet path?  Because if so, that's the dumbest thing I've heard this week, and given how much the current POTUS is in the news, that's saying something.

> * interactive light-field experiences

I have absolutely no clue what practical application you might be referring to in this category.  But I suspect it basically falls into the same set of requirements as the first two items above, just with a different set of input and output devices.

I think I'll leave Paul and Sebastian to comment further in this vein.

 - Jonathan Morton