<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.6000.16674" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=Arial size=2><BR>
<DIV class=gmail_quote>
<DIV class=Ih2E3d>On Wed, Jul 9, 2008 at 9:41 PM, Lachlan Andrew &lt;<A 
href="mailto:lachlan.andrew@gmail.com" 
target=_blank>lachlan.andrew@gmail.com</A>&gt; wrote:<BR>
<BLOCKQUOTE class=gmail_quote 
style="PADDING-LEFT: 1ex; MARGIN: 0pt 0pt 0pt 0.8ex; BORDER-LEFT: rgb(204,204,204) 1px solid">Greetings 
  Adam,<BR><BR>2008/7/9 adam maxiaodong &lt;<A 
  href="mailto:adam.maxiaodong@gmail.com" 
  target=_blank>adam.maxiaodong@gmail.com</A>&gt;:<BR>
  <DIV>&gt;<BR>&gt; But why fairness and friendliness are equally 
  important?<BR>&gt;<BR>&gt; I though all those congestion algorithms are 
  already intra-protocol fair<BR>&gt; (otherwise, it will be a design flaw and 
  won't get implemented)<BR><BR></DIV>Even Reno isn't intra-protocol fair 
  (unless RTTs are equal).<BR>Algorithms get implemented in Linux because 
  someone takes the effort<BR>to implement them, not because they have been 
  shown to be good.<BR><BR>Lars is absolutely right that assessing based only on 
  throughput is<BR>missing the point of congestion control.</BLOCKQUOTE></DIV>
<DIV><BR>what about measuring queuing delays? it would be another important 
metric!<BR><BR>best<BR>saverio<BR></DIV></DIV></FONT></DIV></BODY></HTML>