Message-Id: <200303172317.h2HNHjRq017588@picard2.cs.ohiou.edu> From: "Shawn Ostermann" <ostermann@eecs.ohiou.edu> Subject: Re: tcptrace Slow Start Duration Date: Mon, 17 Mar 2003 18:17:45 -0500
Mark Allman <mallman@grc.nasa.gov> wrote:
> I think the heuristic is easier, actually...
>
> * Watch for a retransmit, which should signal the end of slow
> start.
Unless you become window limited first. Which is still common in my
experience. Whether that's true or not, we get to your next point...
> * Watch for the connection to become window limited (i.e., the
> cwnd == advertised window).
Yes. I'm concerned that this heuristic isn't as obvious as I would
like. Making sure that the observed window is within a MSS of the
advertisement would be a start. I'm concerned that in many traces that
I've seen, the tracing point (while "near" the sender) is just enough
farther away that it's clear in the traces that the connection has
become window limited even though we're technically not near enough the
advertised window.
> I suppose that in the end I am pretty dubious of
> adding these sorts of things to tcptrace
I tend to agree. There's an even bigger heuristic smell to this than
normal, and that makes me uncomfortable.
That being said, however, I've been toying with a heuristic of trying to
characterize a connection into:
1) window limited
(as above)
2) loss-limited
(high amount of retransmission)
3) normal
(or can't tell?)
4) data limited
(not enough application data)
This would answer a common question among many people, "why is this
connection so slow"?
My first though is to characterize, separately, each RTT window and
group each flight into one of the above. Then to see if there's enough
agreement between the intervals to declare an overall behavior.
Suggestions are welcome!
Shawn
-------------------------------------------------------------------------
Dr. Shawn Ostermann - Associate Professor - Ohio University
322B Stocker Center, Ohio University, Athens, Ohio 45701-2979
ostermann@eecs.ohiou.edu -- FAX: (740)593-0007 -- Voice: (740)593-1234
http://ace.cs.ohiou.edu/~osterman http://irg.cs.ohiou.edu
----------------------------------------------------------------------------
To unsubscribe, send a message with body containing "unsubscribe tcptrace" to
majordomo@tcptrace.org.
This archive was generated by hypermail 2b30 : 03/18/03 EST