Re: [tcpPrague] Volunteers pls: L4S non-WG-forming BoF proposal cut-off approaching

[Michael Welzl <[e-mail]>]

> On 22. mai 2016, at 21.17, John Leslie <[e-mail]> wrote:
> 
> Michael Welzl <[e-mail]> wrote:
>> (snip)
>> 
>>>> {Note 1} Some of the deliverables would be brought to tcpm anyway, 
>>>> because they are important for all TCP-derived congestion controls, not 
>>>> just "TCP Prague".
>>>> For instance, the following two RTT-related items are particularly 
>>>> important for L4S/"TCP Prague", but they are also important updates to 
>>>> the base TCP cc [RFC5681] too.
>>>> 3-4) Scaling TCP's Congestion Window for Small Round Trip Times
>>>> 3-5) Reduce TCP / SCTCP / TCP Prague RTT-dependence
>>> 
>>> This is almost off-topic. :^(
>>> 
>>> The "small round-trip times" issue is pretty fundamental: TCP _doesn't_
>>> reduce its congestion window below 2 (packets); so with enough classic-TCP
>>> senders, no AQM can possibly hold latency low. Bob has proposed a change;
>>> but it's a change to _every_ classic TCP sender. Myself, I'd rather not
>>> worry about this just yet...
>>> 
>>> RTT-dependence is an inseparable part of TCP-as-we-know-it. I'd _much_
>>> rather not tilt at that windmill quite yet???
>> 
>> So here???s the +1.  I can???t see how it helps this work if we discuss
>> all the possible problems of the world and how it can possibly solve them.
> 
>   That's a trifle unfair to Bob:

Wasn’t the intention => au contraire…


> he included a bunch of things which
> are actually prerequsites to L4S working well.
> 
>   I don't disagree with Bob that these are important: I merely don't
> want to tie ourselves in knots trying to solve them at the outset.
> 
>> Opening many cans of various worms - wouldn???t it be better to keep
>> things more focused?
> 
>   Exactly!

… that was my point.

Cheers,
MIchael


> 
>   I listed the things I think are necessary to show sufficient results
> to get folks believing the goal is worth the effort. I'd like to concentrate
> our efforts there; and get around-to the other details after we can show
> some big-I-Internet results.
> 
> --
> John Leslie <[e-mail]>