London software testing news UK

Network Performance Testing

Posted in Software testing by testing in London on October 4, 2007

From EDN
So am I correctly interpreting from the documentation that you recommend > leaving Iperf’s buffer length setting at the default 8 KBytes, but explicitly setting the TCP window size to 128 KBytes?

That’s correct. In my tests, the parameter that has most impact in performance is TCP window size. This is due to the fact that the “bandwidth x latency” product in powerline networks is higher than in Fast Ethernet or Gigabit Ethernet networks, so unless you have a large TCP window, your PC won’t be able to fill the pipe with enough data. End users can change that in their computers by changing a registry value. The good thing is that Windows Vista includes auto TCP Window setting. I have not tested it yet, though.

Please don’t take offense at what I’m about to ask, but…do these settings reflect typical real-life configurations, or are they just optimal for DS2’s hardware design and communications protocol (and conversely, perhaps, sub-optimal for HomePlug AV or HD-PLC)?

As far as I know, based on my tests, having a large TCP window is good for every vendor. All powerline technologies have large “bandwidth x latency” products, because of the extra latency created by multiple-access, packet fragmentation and re-assembly, packet retransmissions, etc.

Stress testing services

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: