<span></span>Hi All,<br><br>I don't think I have the email that started this thread, so I might be<br>out of line. But I wanted to suggest that there are some interesting<br>things that happen in slow start on an uncongested system, depending
<br>on the size of the buffer at the bottleneck relative to the bandwidth-delay<br>product. <br><br>With apologies for this shameless act of self-promotion, I have a paper on<br>this topic that you can download here:
<br><br><a href="http://allendowney.com/research/tcp/downey07tcp.pdf" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://allendowney.com/research/tcp/downey07tcp.pdf</a><br><br>If that's useful, let me know. If not, I'm sorry for jumping into the middle!
<br>
<br>Cheers,<br><span class="sg">Allen</span><div><br><span class="q" id="q_11616dee03a25fac_2"><br><br><div><span class="gmail_quote"></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div bgcolor="#ffffff">
<div><span class="gmail_quote">On 11/6/07, <b class="gmail_sendername">Sally
Floyd</b> <<a href="mailto:sallyfloyd@mac.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">sallyfloyd@mac.com</a>> wrote:</span>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">>
- The rise time of a single flow to an empty systems is not very<br>>
interesting, because it many measures the impact of slow
start.<br><br>Actually, the flow completion time in an uncongested system can
be<br>a quite interesting thing to measure, particularly if one
is<br>evaluating one of the many proposals for start-ups faster
than<br>slow-start. (One of these proposals is Quick-Start, RFC
4782; some<br>of the others are discussed in Appendix A of RFC 4782.)<br><br>I
would recommend having scenarios include the case of
a<br>generally-uncongested link, as well as including cases with
various<br>levels of congestion, with metrics including per-flow
transfer<br>times, fairness, and aggregate packet drop rates. This
should<br>illustrate some of the good and potentially-bad aspects of
protocols<br>with fast start-ups.<br><br>-
Sally<br><a href="http://www.icir.org/floyd/" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://www.icir.org/floyd/</a><br><br>RFC
4782: <a href="http://www.ietf.org/rfc/rfc4782.txt" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://www.ietf.org/rfc/rfc4782.txt</a><br></blockquote></div></div></blockquote></div></span>
</div>