I am very happy to see the demand for cake and fq_codel here, and do hope the ipv6 problem is resolved soon. I’ve been trying to help
with cake specific configuration primarily over here: http://forum.mikrotik.com/t/some-quick-comments-on-configuring-cake/152505/1 and here: http://forum.mikrotik.com/t/for-isps-motivations-and-methods-for-implementing-fq-codel-and-cake/154153/1
I am a big fan of the flent.org testing tool (the tcp_nup, rrul, and rtt_fair tests primarily), and there a lot of usage of that and examples on those threads. I hope the mikrotik folk start using it also to drive loads and optimize cpu usage, among other additional things.
We still have not validated the diffserv and ecn support actually work, and I hope someone gets on that. You can also specify -6 or -4 to select ipv6 or ipv4 as the transport. I keep hoping for a 2 ports into one test in particular.
(well, my big hope is that more make fq_codel the default on all interfaces, per https://datatracker.ietf.org/doc/rfc7567/ as it’s the default nowadays for most other linux, openwrt, ios, and osx. It’s really lightweight when run at line rate and way better than a fifo)
It’s also not clear to me if mikrotik picked up the wifi work on any of their wifi gear: https://lwn.net/Articles/705884/
I don’t actually have any mikrotik hardware at this point, and plan to just haunt those two threads for now, although I’d like to find something that can drive multiple 10GigE ports eventually. My biggest feature request of mikrotik is that somehow they start making packets, drops, marks, and reschedules statistics available, and/or any of the more detailed stats cake provides.