[ns] Wireless Simulation, TCP throughput strange results

[Joe Andersson <joe_se01@yahoo.com>]

Hello,

I am simulating a simple wireless network, using DSR,
CBR over TCP sending from node 1 to node 6.

I need to plot the throughput at node 6; I would
expect some kind of constant bitrate, but I get
something that oscillates all the time rapidly from
high throughput to almost 0.

Could you please help me on finding the reasons for
that anomalous behaviour?

The method I use for calculating the throughput is as
follows (give me comments on it, please):
I trace everything, I filter out the correctly
received packets at node 6 and from the clocks of
these events I estimate the throughput at that time as
bits_per_packet /(clock_now -
clock_last_received_packet).

Any input is much appreciated,

Thanks
/J


__________________________________________________
Terrorist Attacks on U.S. - How can you help?
Donate cash, emergency relief information
http://dailynews.yahoo.com/fc/US/Emergency_Information/