PTP 650 - Bandwidth difference.

Dear, I need help.

I have an PTP650 Integrated installed at a distance of approximately 500 meters.

It turns out that when I test in UDP, it passes 221 / 222mbps simultaneously. When I test in TCP, it gives a peak of 180mbps and then drops and stabilizes at 50 / 60mbps.


TCP Test - Both Sides simultaneously

UDP Test - Both Sides simultaneously

PTP 650 Wireless.png

What could be happening? I need more TCP bandwidth at this location.


I know that the UDP performance isn't your main problem, but let's look at that for a moment. The test tool is reporting 221 Mbit/s and 222 Mbit/s. It's useful to know what the tool is really measuring. For example, is it measuring capacity for Ethernet frames, IP datagrams or UDP messages? If the test tool is measuring capacity for Ethernet frames then the measurement is a slightly lower than the actual capacity of the link, shown as 226 Mbit/s in the PTP 650 user interface. That might indicate a small error in the measurements.

If the test tool is measuring the capacity for IP datagrams then this is equivalent to 224 Mbit/s for Ethernet traffic in the link. If the test tool is measuring only the UDP messages, then the measurement is equivalent to 226 Mbit/s in the link, which is exactly the calculated link capacity.

You also show some results for TCP, and I guess you're disappointed that the measured result is significantly lower than the displayed link capacity for the PTP 650. The PTP 650 link is transparent to the IP, UDP or TCP protocols in your traffic, and the underlying link capacity will be exactly the same in every case.

The lower TCP capacity reported in your test tool reflects how efficiently or inefficiently the TCP end points are using the capacity of the link. In other words, how efficiently they are saturating the link. This efficiency can be affected by the window size and other configuration of the TCP function in the test system. If the tester is executing on a PC, it may also reflect the underlying performance of the PC. A sustained biderectional transfer of more than 200 Mbit/s is likely to load most PCs quite heavily.

You can check the performance of the PC by connecting the testers without a wireless link. If this returns a result of much more than 200 Mbit/s in each direction then your test system is good enough to measure the capacity of the link. If the test system has good performance, please review the window size and other settings for TCP in the tester.

Please keep us up to date if you get any further results.


1 Like

Hi Mark, thank you for the explanation!

The tool I'm using is Dude (Bandwidth Test inside), but, i do a new band test using the direct test tool from one of my mikrotik Cloud Core Routers - that I own - for the other and the rate is completely different see:

I think the problem is actually the test tool I used earlier.