Re: cwnd bug?

[Christof Brandauer <cbrand@cosy.sbg.ac.at>]

On Tue, 25 May 1999, Andras Veres wrote:

: I tried the following little script to log the cwnd of two competing
: TCPs. They share a single buffer:
: 
: $ns duplex-link $na $nb 0.2Mb 10ms DropTail
: 
: Surprisingly the cwnd increases continuously which is quite
: strange. Even after simulating it for 1 hour simulation time.
: 
: Is it a bug or a 'feature' :) ?

It's a feature. Although there is no dynamic window advertisement in
ns there is one variable that sets an upper bound on a TCP sender's
window size:

Agent/TCP set window_ 20   (see ~ns/tcl/lib/ns-default.tcl)

The usable window = min (cwnd_, window_)

-> in your sims the effectively used window is determined by window_
(once cwnd_ > window_).

-> with this window size the queue sizze converges to ~40 pkts
-> no packet loss
-> cwnd_ keeps growing... (slow start/cong. avoidance)


in your sim script you can add something like:

>       set str "$str [$tcp($j) set cwnd_ ] [$tcp($j) set window_]"

to see which variable controls the send window.


hope this helps,
chris


 -----------------------------------------------------------------
  Christof Brandauer                        cbrand@cosy.sbg.ac.at
  Student of Applied Informatics                 
  Department of Computer Science, University of Salzburg, Austria
 -----------------------------------------------------------------