<html><head><style type="text/css"><!-- DIV {margin:0px;} --></style></head><body><div style="font-family:times new roman, new york, times, serif;font-size:10pt"><DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif">Lachlan,</DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif"> </DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif">I dont see a problem in measuring transient interaction - I see an opportunity. If a researcher wishes to evaluate the transient of a single flow, he/she can certainly design an experiment to make sure interactions are avoided. (spacing flows properly, as you mentioned). I find transient interaction effects VERY interesting to study...</DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif"> </DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif">The bottom line is: agreeing on X(t) performance metrics instead of x(t) does not LIMIT in any way the expressiveness of experimental results, since we can always reduce X(t) to x(t) by using a single flow...</DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif"> </DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif">Dirceu<BR><BR></DIV>
<DIV style="FONT-SIZE: 12pt; FONT-FAMILY: times new roman, new york, times, serif">----- Original Message ----<BR>From: Lachlan Andrew <lachlan.andrew@gmail.com><BR>To: Xiaoliang (David) Wei <weixl@caltech.edu><BR>Cc: tmrg <tmrg-interest@icsi.berkeley.edu>; iccrg IRTF list <iccrg@cs.ucl.ac.uk><BR>Sent: Wednesday, October 31, 2007 10:18:13 AM<BR>Subject: [Iccrg] Re: [Tmrg] convergence time<BR><BR>Greetings David,<BR><BR>On 28/10/2007, Xiaoliang (David) Wei <<A href="mailto:weixl@caltech.edu" ymailto="mailto:weixl@caltech.edu">weixl@caltech.edu</A>> wrote:<BR>> Another option, to eliminate the dependency to stability and<BR>> timescale, is that we don't study the convergence of current rate.<BR>> Instead, we study the convergence of the aggregate average rate. That is,<BR>> if the instanenous rate of a flow at time t is x(t), we define the<BR>> aggregate average rate of the flow at time t to
be<BR>> X(t) = 1/t * sum u=0->t x(u).<BR>> ("sum" can be "integrate" if the time is continous).<BR><BR>Good idea.<BR><BR>> Then we study the convergence of the curve X(t) to the "final value".<BR>> This process might be easier as:<BR>> 1. X(t) is easier to measure because we can just look at the amount we<BR>> have transfered from time 0 to time t;<BR>> 2. X(t) converges even x(t) has a limit-cycle oscillation, so it is less<BR>> sensitive to stability<BR>> 3. If x(t) converges fast, X(t) converges fast too. We can still compare<BR>> the convergence with X(t)<BR>> 4. X(t) does have meaning in user-experience. It measures how long the<BR>> users have to participate in the network to get to the desired rate.<BR><BR>They're all good points.<BR><BR>The main drawback is that X(t) converges (much) more slowly, since<BR>it always gives some weight to the early rates. If we want to observe<BR>the impact
of each of several newly arriving flows, we need to space<BR>them out further if we use X(t) than we do if we use x(t), or else<BR>the transients will interact.<BR><BR>The time required to find the "final" value could already be quite<BR>long, especially in the case of Reno, which takes hours to reach<BR>equilibrium on large BDP paths.<BR><BR>What do others think?<BR><BR>Cheers,<BR>Lachlan<BR><BR>-- <BR>Lachlan Andrew Dept of Computer Science, Caltech<BR>1200 E California Blvd, Mail Code 256-80, Pasadena CA 91125, USA<BR>Ph: +1 (626) 395-8820 Fax: +1 (626) 568-3603<BR><A href="http://netlab.caltech.edu/~lachlan" target=_blank>http://netlab.caltech.edu/~lachlan</A><BR><BR>_______________________________________________<BR>Iccrg mailing list<BR><A href="mailto:Iccrg@cs.ucl.ac.uk" ymailto="mailto:Iccrg@cs.ucl.ac.uk">Iccrg@cs.ucl.ac.uk</A><BR><A href="http://oakham.cs.ucl.ac.uk/mailman/listinfo/iccrg"
target=_blank>http://oakham.cs.ucl.ac.uk/mailman/listinfo/iccrg</A><BR></DIV>
<DIV style="FONT-SIZE: 10pt; FONT-FAMILY: times new roman, new york, times, serif"><BR></DIV></div><br>__________________________________________________<br>Do You Yahoo!?<br>Tired of spam? Yahoo! Mail has the best spam protection around <br>http://mail.yahoo.com </body></html>