On Wed, 1 Sep 2010, [email protected] wrote: > I currently have Iperf installed on Windows XP machines and I'm > curious about the behavior of the tool, but more specifically, the -w switch. > [...] > This is my understanding so far, please correct me if (read: when) I'm > wrong: > > (1) The -w switch really isn't the TCP Receive Window as the > documentation says but rather the send/receive socket buffer size, and > is not directly related to TCP Window Size. This goes against what I had > originally assumed from the documentation and pretty much everything > I've read online.
You correctly understood that the badly misnamed -w switch is merely sending a polite socket buffer size suggestion to the operating system. The final buffer size decided by the operating system is *indirectly* related to the TCP window size, but *only on the TCP receiver* and in no other case. Part of the confusion comes from the fact that increasing the *send* socket buffer size is ALSO beneficial, for a reason similar yet totally unrelated to the (receive) TCP window: <http://en.wikipedia.org/wiki/TCP_tuning#Window_size> -------- The more general problem is that people give iperf way too much credit (sorry iperf developers, I do not mean this in a bad way). All iperf does is stressing the operating system (and it is good at it). TCP (and UDP FWIW) are entirely implemented in your operating system; there is not the tiniest bit of TCP to be found in iperf. Your operating system, NOT iperf, is entirely responsible of every single network performance aspect. In other words: no, "i-Perf" does not "perf". Things will probably get better once the iperf documentation finally fixes the documentation of the -w switch and stops mentioning the TCP window (= the cheapest and most accurate way to stop spreading lies and confusion about it). So, whenever you study network performance, do NOT use "iperf" as a google keyword or you will miss all the best information, finding only discussions of clueless people who wrongly believe iperf is somehow in charge of network performance (minor exaggeration here to get the point across) > (4) Should the -w switch be used for UDP tests? Typically yes, but again it really depends on the UDP implementation in your operating system. ------------------------------------------------------------------------------ This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd _______________________________________________ Iperf-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/iperf-users
