test here
http://www.dslreports.com/speedtest click the green dsl button to start, or fibre if you want, fibre is same test but with more threads , it measures latency during speed tests to check for buffer bloat. My results are with the dsl test tho.
So here is first test on my old router, no QoS, CTF acceleration enabled.
http://www.dslreports.com/speedtest/419540Here is 2nd test after I got a router powerful enough to disable CTF and use QoS, this is with upstream QoS enabled, but due to a bug in asus firmware, no downstream QoS.
http://www.dslreports.com/speedtest/681469Here is a 3rd test after I manually configured some ingress QoS, (I have notified dev of firmware of fix).
http://www.dslreports.com/speedtest/713055The penalty for this is capped throughput rates, the upload was reduced by 10%, it may work fine higher tho which I will experiment with, I only reduced the downstream by a much smaller amount tho. Also single threaded takes a larger hit as tcp is more conservative in growing its tcp buffers.
TBB test here.
So can see single threaded doesnt aggressively max its buffers and as such is slower.
This is much more effective than what plusnet's isp side shaping achieves. The first result is when left to plusnet QoS alone.