[Cake] [rmcat] [rtcweb] Catching up on diffserv markings

[Sebastian Moeller]

Hi Justin,

On Oct 27, 2015, at 17:16 , Justin Uberti <juberti at google.com> wrote:

> 
> 
> On Tue, Oct 27, 2015 at 1:06 AM, Sebastian Moeller <moeller0 at gmx.de> wrote:
> Hi Justin,
> 
> 
> On Oct 22, 2015, at 21:54 , Justin Uberti <juberti at google.com> wrote:
> 
> > At present I'm not aware of any widely-deployed OS where an app can read the received ECN markings.
> >
> > iOS9 added support for this within the kernel, and it's used for TCP, but not exposed to userspace. There is an open Radar bug asking for this info to be exposed to userspace.
> >
> > FWIW, Chrome supports setting the DSCP markings if you set a magic parameter. But it's not on by default, mainly because we've never done the auditing necessary to ensure this doesn't randomly break in various dimly-lit parts of the internet.
> 
>         Slightly related question, is this DSCP marking capability restricted to webrtc packets or is there a way to make chrome use (arbitrary) DSCP marks for all its packets? For exercising different priority banding schemes such an option would be perfect (say to test whether a aqm+qos system will allow snappy browsing even with heavy download/upload/bittorrent traffic in other priority bands; this test is especially interesting if all traffic sources can reside on the same host, as this is a quite common set-up in home networks, one computer that does everything concurrently and where the users still want a decent browsing experience).
> 
> The option that currently exists only works for WebRTC packets. 

Ah, thanks, that is a pity. It would be really sweet for testing different qos systems. I guess on a linux machine I should e able to fake it with a local iptable rule…

Best Regards
	Sebastian