I have some HP ProCurve series Switch. There is a QoS option for prioritize traffic by TCP/UDP port. I set two ports: one with highest priority and another with lowest. Then I performed tests to ensure that prioritizing really works. I run two instances (per computer) of Network Benchmark application on two computers being in the LAN. One computer acts like server and other like client. In benchmark applications I set two ports configured earlier on the switch. Then I run tests, so that server sends some amount of data to client showing current transmission speed. I assumed that, since one port has higher priority then other, the transmission speed will not be equal, and the traffic going throughout the tcp port with higher priority will be significant faster. Unfortunately there is no difference. The speeds are nearly equal (half of bandwidth) and non prioritizing seems to be applied.
Can anyone explain me why this is not working. I suppose that I misunderstand some QoS basics and this is a reason, but I will be grateful if someone clarify this issue for me.