<div class="gmail_quote">On Fri, Aug 12, 2011 at 12:48, rick jones <span dir="ltr"><<a href="mailto:perfgeek@mac.com">perfgeek@mac.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im"><br>
On Aug 12, 2011, at 7:03 AM, Barry Constantine wrote:<br>
<br>
> Hi,<br>
><br>
> I did some testing to compare various TCP stack behaviors in the midst of traffic policing.<br>
><br>
> It is common practice for a network provider to police traffic to a subscriber level agreement (SLA).<br>
><br>
> In the iperf testing I conducted, the following set-up was used:<br>
><br>
> Client -> Delay (50ms RTT) -> Cisco (with 10M Policing) -> Server<br>
><br>
> The delay was induced using hardware base commercial gear.<br>
><br>
> 50 msec RTT and bottleneck bandwidth = 10 Mbps, so BDP was 62,000 bytes.<br>
><br>
> Ran Linux, Windows XP, and Windows 7 clients at 32k, 64k, 128k window (knowing that policing would<br>
> kick in at 64K)<br>
><br>
> Throughput for Window (Mbps)<br>
><br>
> Platform 32K 64K 128K<br>
> --------------------------------------------<br>
> Linux 4.9 7.5 3.8<br>
> XP 5.8 6.6 5.2<br>
> Win7 5.3 3.4 0.44<br>
><br>
<br>
</div>The folks in tcpm might be better able to help? but I'll point-out one nit - "Linux" is not that much more specific than saying "Unix" - it would be goodness to get into the habit of including the kernel version. And ID the server since it takes two to TCP...<br>
</blockquote><div><br></div><div>BTW, including latest FreeBSD (or some other BSD) and Mac OS X might give interesting - maybe useful - values for the work.</div><div><br></div><div>Alex </div></div>