[ns] Suggestions in modifying 802.11 mac
andy80 at gmail.com
Tue Sep 20 22:35:52 PDT 2005
Finally, i have manged to get my simulation done.
What my simluation does is that i have setup 7 nodes, all 200 metres apart
each and all connected wire wireless 802.11. Adhoc routing is DSR. There are
2 tcp connections, tcp1 and tcp2. Source of tcp1 is at node 0, and tcpSink1
is at node 4. Source of tcp2 is at node 5 and tcpSink2 is at node 6. 2 FTP
sources are also created for each of the tcp connections respectively. The
simulation time is set at 10. ftp1 will start at time 1 and end at time 10.
ftp2 will start at time 5 and end at time 8. Simulation runs fine and i have
got my trace files.
I plotted the throughput graphs with tracegraph (wonderful tool!) and my
graph shows that when ftp2 starts to send at time 5, ftp1 will be "switched"
off and when ftp2 has finished sending, ftp1 will resume and finish at time
10. in other words, i have shown that 2 tcp connection cannot co-exist.
The next part of my experiment is to attempt to modify the mac of 802.11 to
change the backoff time so as to allow both tcp connections to co-exist. Any
suggestions on how to approach this?
Cheers, thanks everyone for all the help which made my simulation possible,
Would you like a Google Mail account with 1GB of email space?
More information about the Ns-users