xnor
Occasional Visitor
The simple_shaper.sh should work fine with the RT-N56U, but I planned on adding a more complex script that also allows for fair download speed when multiple people are sharing the Internet connection. This does not work with the original builds, since they don't include the necessary kernel modules.can it also been used for ASUS rt-n56u?
If there's interest I could also compile RT-N56U images, but I don't have the hardware to test them.
I guess you're interested in the "Network buffer measurements". Well, with the shaping script you can reduce uplink buffer time to almost 0.what are your results on the netalyzr site?
Only thing you cannot influence directly is the ISP side (downlink). Everything but the downlink buffer measurement is green when I reduce the buffers in my script and run the test.
BUT, and this is a big but, this test doesn't tell you what's really going on.
I can set buffers high for bulk transfers but still prioritize e.g. voice or game traffic so that it will be sent without delay. My script doesn't just do simple shaping but also prioritization.
The netalyzr seems to send UDP packets as fast as possible, which is way way faster than the uplink can handle. No real application does that. It seems to measure the time until the buffer is full and the router/modem starts dropping all new packets.
Real-world applications don't send data as fast as possible, they either wait for replies or acknowledgments or in case of TCP transmissions simply get blocked by the operating system.
There may however be a problem with applications that send bursts of traffic when you make the buffer very small. If a traffic burst fills up the very short queue the next packets will simply be dropped by the router. A voice chat or streaming application could experience horrible packet loss.
Last edited: