On Nov 27, 2017, at 3:07 PM, Pete Heist <peteheist@gmail.com> wrote:

On Nov 27, 2017, at 3:01 PM, Jonathan Morton <chromatix99@gmail.com> wrote:

Looking at the Cake stats for that run, it doesn't seem to have been signalling congestion at all, when you'd expect it to with 13 bulk flows running through it.  Something odd is going on there.

Ok, I’ll re-run this at a lower rate to make sure I’m not running out of CPU, which I suppose would be more likely with the dual-whatever keywords than not.

That’s almost for sure the problem, as fairness works as expected at 500mbit instead of 950mbit:

http://drhleny.cz/bufferbloat/cake/round1/hostiso_eg_cake_dsrc_cake_ddst_500mbit/index.html

Either I have to test at this reduced rate or use fewer flows for this test. The netperf instances were probably not able to output in total at line rate, and going from srchost/dsthost to dual-srchost/dual-dsthost was probably enough to make the difference in CPU consumption.

Follow-up question, in theory, would it be possible for cake to know that it doesn’t have enough CPU to operate properly so it can emit a warning every so often?